# HG changeset patch # User davidvanzessen # Date 1548752049 18000 # Node ID ba33b94637ca874848b4a9019f5ac6d968c65bbc # Parent 43a1aa648537827ab031b667d2ade4ec8b4fc7a0 Uploaded diff -r 43a1aa648537 -r ba33b94637ca LICENSE --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/LICENSE Tue Jan 29 03:54:09 2019 -0500 @@ -0,0 +1,21 @@ +MIT License + +Copyright (c) 2019 david + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. \ No newline at end of file diff -r 43a1aa648537 -r ba33b94637ca README.md --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/README.md Tue Jan 29 03:54:09 2019 -0500 @@ -0,0 +1,12 @@ +# SHM CSR + +Somatic hypermutation and class switch recombination pipeline + +# Dependencies +-------------------- +[Python 2.7](https://www.python.org/) +[Change-O](https://changeo.readthedocs.io/en/version-0.4.4/) +[Baseline](http://selection.med.yale.edu/baseline/) +[R data.table](https://cran.r-project.org/web/packages/data.table/data.table.pdf) +[R ggplot2](https://cran.r-project.org/web/packages/ggplot2/ggplot2.pdf) +[R reshape2](https://cran.r-project.org/web/packages/reshape/reshape.pdf) \ No newline at end of file diff -r 43a1aa648537 -r ba33b94637ca aa_histogram.r --- a/aa_histogram.r Thu Dec 07 03:44:38 2017 -0500 +++ b/aa_histogram.r Tue Jan 29 03:54:09 2019 -0500 @@ -1,69 +1,69 @@ -library(ggplot2) - -args <- commandArgs(trailingOnly = TRUE) - -mutations.by.id.file = args[1] -absent.aa.by.id.file = args[2] -genes = strsplit(args[3], ",")[[1]] -genes = c(genes, "") -outdir = args[4] - - -print("---------------- read input ----------------") - -mutations.by.id = read.table(mutations.by.id.file, sep="\t", fill=T, header=T, quote="") -absent.aa.by.id = read.table(absent.aa.by.id.file, sep="\t", fill=T, header=T, quote="") - -for(gene in genes){ - graph.title = paste(gene, "AA mutation frequency") - if(gene == ""){ - mutations.by.id.gene = mutations.by.id[!grepl("unmatched", mutations.by.id$best_match),] - absent.aa.by.id.gene = absent.aa.by.id[!grepl("unmatched", absent.aa.by.id$best_match),] - - graph.title = "AA mutation frequency all" - } else { - mutations.by.id.gene = mutations.by.id[grepl(paste("^", gene, sep=""), mutations.by.id$best_match),] - absent.aa.by.id.gene = absent.aa.by.id[grepl(paste("^", gene, sep=""), absent.aa.by.id$best_match),] - } - print(paste("nrow", gene, nrow(absent.aa.by.id.gene))) - if(nrow(mutations.by.id.gene) == 0){ - next - } - - mutations.at.position = colSums(mutations.by.id.gene[,-c(1,2)]) - aa.at.position = colSums(absent.aa.by.id.gene[,-c(1,2,3,4)]) - - dat_freq = mutations.at.position / aa.at.position - dat_freq[is.na(dat_freq)] = 0 - dat_dt = data.frame(i=1:length(dat_freq), freq=dat_freq) - - - print("---------------- plot ----------------") - - m = ggplot(dat_dt, aes(x=i, y=freq)) + theme(axis.text.x = element_text(angle = 90, hjust = 1), text = element_text(size=13, colour="black")) - m = m + geom_bar(stat="identity", colour = "black", fill = "darkgrey", alpha=0.8) + scale_x_continuous(breaks=dat_dt$i, labels=dat_dt$i) - m = m + annotate("segment", x = 0.5, y = -0.05, xend=26.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 13, y = -0.1, label="FR1") - m = m + annotate("segment", x = 26.5, y = -0.07, xend=38.5, yend=-0.07, colour="darkblue", size=1) + annotate("text", x = 32.5, y = -0.15, label="CDR1") - m = m + annotate("segment", x = 38.5, y = -0.05, xend=55.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 47, y = -0.1, label="FR2") - m = m + annotate("segment", x = 55.5, y = -0.07, xend=65.5, yend=-0.07, colour="darkblue", size=1) + annotate("text", x = 60.5, y = -0.15, label="CDR2") - m = m + annotate("segment", x = 65.5, y = -0.05, xend=104.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 85, y = -0.1, label="FR3") - m = m + expand_limits(y=c(-0.1,1)) + xlab("AA position") + ylab("Frequency") + ggtitle(graph.title) - m = m + theme(panel.background = element_rect(fill = "white", colour="black"), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank()) - #m = m + scale_colour_manual(values=c("black")) - - print("---------------- write/print ----------------") - - - dat.sums = data.frame(index=1:length(mutations.at.position), mutations.at.position=mutations.at.position, aa.at.position=aa.at.position) - - write.table(dat.sums, paste(outdir, "/aa_histogram_sum_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) - write.table(mutations.by.id.gene, paste(outdir, "/aa_histogram_count_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) - write.table(absent.aa.by.id.gene, paste(outdir, "/aa_histogram_absent_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) - write.table(dat_dt, paste(outdir, "/aa_histogram_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) - - png(filename=paste(outdir, "/aa_histogram_", gene, ".png", sep=""), width=1280, height=720) - print(m) - dev.off() - - ggsave(paste(outdir, "/aa_histogram_", gene, ".pdf", sep=""), m, width=14, height=7) -} +library(ggplot2) + +args <- commandArgs(trailingOnly = TRUE) + +mutations.by.id.file = args[1] +absent.aa.by.id.file = args[2] +genes = strsplit(args[3], ",")[[1]] +genes = c(genes, "") +outdir = args[4] + + +print("---------------- read input ----------------") + +mutations.by.id = read.table(mutations.by.id.file, sep="\t", fill=T, header=T, quote="") +absent.aa.by.id = read.table(absent.aa.by.id.file, sep="\t", fill=T, header=T, quote="") + +for(gene in genes){ + graph.title = paste(gene, "AA mutation frequency") + if(gene == ""){ + mutations.by.id.gene = mutations.by.id[!grepl("unmatched", mutations.by.id$best_match),] + absent.aa.by.id.gene = absent.aa.by.id[!grepl("unmatched", absent.aa.by.id$best_match),] + + graph.title = "AA mutation frequency all" + } else { + mutations.by.id.gene = mutations.by.id[grepl(paste("^", gene, sep=""), mutations.by.id$best_match),] + absent.aa.by.id.gene = absent.aa.by.id[grepl(paste("^", gene, sep=""), absent.aa.by.id$best_match),] + } + print(paste("nrow", gene, nrow(absent.aa.by.id.gene))) + if(nrow(mutations.by.id.gene) == 0){ + next + } + + mutations.at.position = colSums(mutations.by.id.gene[,-c(1,2)]) + aa.at.position = colSums(absent.aa.by.id.gene[,-c(1,2,3,4)]) + + dat_freq = mutations.at.position / aa.at.position + dat_freq[is.na(dat_freq)] = 0 + dat_dt = data.frame(i=1:length(dat_freq), freq=dat_freq) + + + print("---------------- plot ----------------") + + m = ggplot(dat_dt, aes(x=i, y=freq)) + theme(axis.text.x = element_text(angle = 90, hjust = 1), text = element_text(size=13, colour="black")) + m = m + geom_bar(stat="identity", colour = "black", fill = "darkgrey", alpha=0.8) + scale_x_continuous(breaks=dat_dt$i, labels=dat_dt$i) + m = m + annotate("segment", x = 0.5, y = -0.05, xend=26.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 13, y = -0.1, label="FR1") + m = m + annotate("segment", x = 26.5, y = -0.07, xend=38.5, yend=-0.07, colour="darkblue", size=1) + annotate("text", x = 32.5, y = -0.15, label="CDR1") + m = m + annotate("segment", x = 38.5, y = -0.05, xend=55.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 47, y = -0.1, label="FR2") + m = m + annotate("segment", x = 55.5, y = -0.07, xend=65.5, yend=-0.07, colour="darkblue", size=1) + annotate("text", x = 60.5, y = -0.15, label="CDR2") + m = m + annotate("segment", x = 65.5, y = -0.05, xend=104.5, yend=-0.05, colour="darkgreen", size=1) + annotate("text", x = 85, y = -0.1, label="FR3") + m = m + expand_limits(y=c(-0.1,1)) + xlab("AA position") + ylab("Frequency") + ggtitle(graph.title) + m = m + theme(panel.background = element_rect(fill = "white", colour="black"), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank()) + #m = m + scale_colour_manual(values=c("black")) + + print("---------------- write/print ----------------") + + + dat.sums = data.frame(index=1:length(mutations.at.position), mutations.at.position=mutations.at.position, aa.at.position=aa.at.position) + + write.table(dat.sums, paste(outdir, "/aa_histogram_sum_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) + write.table(mutations.by.id.gene, paste(outdir, "/aa_histogram_count_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) + write.table(absent.aa.by.id.gene, paste(outdir, "/aa_histogram_absent_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) + write.table(dat_dt, paste(outdir, "/aa_histogram_", gene, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T) + + png(filename=paste(outdir, "/aa_histogram_", gene, ".png", sep=""), width=1280, height=720) + print(m) + dev.off() + + ggsave(paste(outdir, "/aa_histogram_", gene, ".pdf", sep=""), m, width=14, height=7) +} diff -r 43a1aa648537 -r ba33b94637ca baseline/Baseline_Functions.r --- a/baseline/Baseline_Functions.r Thu Dec 07 03:44:38 2017 -0500 +++ b/baseline/Baseline_Functions.r Tue Jan 29 03:54:09 2019 -0500 @@ -1,2287 +1,2287 @@ -######################################################################################### -# License Agreement -# -# THIS WORK IS PROVIDED UNDER THE TERMS OF THIS CREATIVE COMMONS PUBLIC LICENSE -# ("CCPL" OR "LICENSE"). THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER -# APPLICABLE LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UNDER THIS LICENSE -# OR COPYRIGHT LAW IS PROHIBITED. -# -# BY EXERCISING ANY RIGHTS TO THE WORK PROVIDED HERE, YOU ACCEPT AND AGREE TO BE -# BOUND BY THE TERMS OF THIS LICENSE. TO THE EXTENT THIS LICENSE MAY BE CONSIDERED -# TO BE A CONTRACT, THE LICENSOR GRANTS YOU THE RIGHTS CONTAINED HERE IN -# CONSIDERATION OF YOUR ACCEPTANCE OF SUCH TERMS AND CONDITIONS. -# -# BASELIne: Bayesian Estimation of Antigen-Driven Selection in Immunoglobulin Sequences -# Coded by: Mohamed Uduman & Gur Yaari -# Copyright 2012 Kleinstein Lab -# Version: 1.3 (01/23/2014) -######################################################################################### - -# Global variables - - FILTER_BY_MUTATIONS = 1000 - - # Nucleotides - NUCLEOTIDES = c("A","C","G","T") - - # Amino Acids - AMINO_ACIDS <- c("F", "F", "L", "L", "S", "S", "S", "S", "Y", "Y", "*", "*", "C", "C", "*", "W", "L", "L", "L", "L", "P", "P", "P", "P", "H", "H", "Q", "Q", "R", "R", "R", "R", "I", "I", "I", "M", "T", "T", "T", "T", "N", "N", "K", "K", "S", "S", "R", "R", "V", "V", "V", "V", "A", "A", "A", "A", "D", "D", "E", "E", "G", "G", "G", "G") - names(AMINO_ACIDS) <- c("TTT", "TTC", "TTA", "TTG", "TCT", "TCC", "TCA", "TCG", "TAT", "TAC", "TAA", "TAG", "TGT", "TGC", "TGA", "TGG", "CTT", "CTC", "CTA", "CTG", "CCT", "CCC", "CCA", "CCG", "CAT", "CAC", "CAA", "CAG", "CGT", "CGC", "CGA", "CGG", "ATT", "ATC", "ATA", "ATG", "ACT", "ACC", "ACA", "ACG", "AAT", "AAC", "AAA", "AAG", "AGT", "AGC", "AGA", "AGG", "GTT", "GTC", "GTA", "GTG", "GCT", "GCC", "GCA", "GCG", "GAT", "GAC", "GAA", "GAG", "GGT", "GGC", "GGA", "GGG") - names(AMINO_ACIDS) <- names(AMINO_ACIDS) - - #Amino Acid Traits - #"*" "A" "C" "D" "E" "F" "G" "H" "I" "K" "L" "M" "N" "P" "Q" "R" "S" "T" "V" "W" "Y" - #B = "Hydrophobic/Burried" N = "Intermediate/Neutral" S="Hydrophilic/Surface") - TRAITS_AMINO_ACIDS_CHOTHIA98 <- c("*","N","B","S","S","B","N","N","B","S","B","B","S","N","S","S","N","N","B","B","N") - names(TRAITS_AMINO_ACIDS_CHOTHIA98) <- sort(unique(AMINO_ACIDS)) - TRAITS_AMINO_ACIDS <- array(NA,21) - - # Codon Table - CODON_TABLE <- as.data.frame(matrix(NA,ncol=64,nrow=12)) - - # Substitution Model: Smith DS et al. 1996 - substitution_Literature_Mouse <- matrix(c(0, 0.156222928, 0.601501588, 0.242275484, 0.172506739, 0, 0.241239892, 0.586253369, 0.54636291, 0.255795364, 0, 0.197841727, 0.290240811, 0.467680608, 0.24207858, 0),nrow=4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) - substitution_Flu_Human <- matrix(c(0,0.2795596,0.5026927,0.2177477,0.1693210,0,0.3264723,0.5042067,0.4983549,0.3328321,0,0.1688130,0.2021079,0.4696077,0.3282844,0),4,4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) - substitution_Flu25_Human <- matrix(c(0,0.2580641,0.5163685,0.2255674,0.1541125,0,0.3210224,0.5248651,0.5239281,0.3101292,0,0.1659427,0.1997207,0.4579444,0.3423350,0),4,4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) - load("FiveS_Substitution.RData") - - # Mutability Models: Shapiro GS et al. 2002 - triMutability_Literature_Human <- matrix(c(0.24, 1.2, 0.96, 0.43, 2.14, 2, 1.11, 1.9, 0.85, 1.83, 2.36, 1.31, 0.82, 0.52, 0.89, 1.33, 1.4, 0.82, 1.83, 0.73, 1.83, 1.62, 1.53, 0.57, 0.92, 0.42, 0.42, 1.47, 3.44, 2.58, 1.18, 0.47, 0.39, 1.12, 1.8, 0.68, 0.47, 2.19, 2.35, 2.19, 1.05, 1.84, 1.26, 0.28, 0.98, 2.37, 0.66, 1.58, 0.67, 0.92, 1.76, 0.83, 0.97, 0.56, 0.75, 0.62, 2.26, 0.62, 0.74, 1.11, 1.16, 0.61, 0.88, 0.67, 0.37, 0.07, 1.08, 0.46, 0.31, 0.94, 0.62, 0.57, 0.29, NA, 1.44, 0.46, 0.69, 0.57, 0.24, 0.37, 1.1, 0.99, 1.39, 0.6, 2.26, 1.24, 1.36, 0.52, 0.33, 0.26, 1.25, 0.37, 0.58, 1.03, 1.2, 0.34, 0.49, 0.33, 2.62, 0.16, 0.4, 0.16, 0.35, 0.75, 1.85, 0.94, 1.61, 0.85, 2.09, 1.39, 0.3, 0.52, 1.33, 0.29, 0.51, 0.26, 0.51, 3.83, 2.01, 0.71, 0.58, 0.62, 1.07, 0.28, 1.2, 0.74, 0.25, 0.59, 1.09, 0.91, 1.36, 0.45, 2.89, 1.27, 3.7, 0.69, 0.28, 0.41, 1.17, 0.56, 0.93, 3.41, 1, 1, NA, 5.9, 0.74, 2.51, 2.24, 2.24, 1.95, 3.32, 2.34, 1.3, 2.3, 1, 0.66, 0.73, 0.93, 0.41, 0.65, 0.89, 0.65, 0.32, NA, 0.43, 0.85, 0.43, 0.31, 0.31, 0.23, 0.29, 0.57, 0.71, 0.48, 0.44, 0.76, 0.51, 1.7, 0.85, 0.74, 2.23, 2.08, 1.16, 0.51, 0.51, 1, 0.5, NA, NA, 0.71, 2.14), nrow=64,byrow=T) - triMutability_Literature_Mouse <- matrix(c(1.31, 1.35, 1.42, 1.18, 2.02, 2.02, 1.02, 1.61, 1.99, 1.42, 2.01, 1.03, 2.02, 0.97, 0.53, 0.71, 1.19, 0.83, 0.96, 0.96, 0, 1.7, 2.22, 0.59, 1.24, 1.07, 0.51, 1.68, 3.36, 3.36, 1.14, 0.29, 0.33, 0.9, 1.11, 0.63, 1.08, 2.07, 2.27, 1.74, 0.22, 1.19, 2.37, 1.15, 1.15, 1.56, 0.81, 0.34, 0.87, 0.79, 2.13, 0.49, 0.85, 0.97, 0.36, 0.82, 0.66, 0.63, 1.15, 0.94, 0.85, 0.25, 0.93, 1.19, 0.4, 0.2, 0.44, 0.44, 0.88, 1.06, 0.77, 0.39, 0, 0, 0, 0, 0, 0, 0.43, 0.43, 0.86, 0.59, 0.59, 0, 1.18, 0.86, 2.9, 1.66, 0.4, 0.2, 1.54, 0.43, 0.69, 1.71, 0.68, 0.55, 0.91, 0.7, 1.71, 0.09, 0.27, 0.63, 0.2, 0.45, 1.01, 1.63, 0.96, 1.48, 2.18, 1.2, 1.31, 0.66, 2.13, 0.49, 0, 0, 0, 2.97, 2.8, 0.79, 0.4, 0.5, 0.4, 0.11, 1.68, 0.42, 0.13, 0.44, 0.93, 0.71, 1.11, 1.19, 2.71, 1.08, 3.43, 0.4, 0.67, 0.47, 1.02, 0.14, 1.56, 1.98, 0.53, 0.33, 0.63, 2.06, 1.77, 1.46, 3.74, 2.93, 2.1, 2.18, 0.78, 0.73, 2.93, 0.63, 0.57, 0.17, 0.85, 0.52, 0.31, 0.31, 0, 0, 0.51, 0.29, 0.83, 0.54, 0.28, 0.47, 0.9, 0.99, 1.24, 2.47, 0.73, 0.23, 1.13, 0.24, 2.12, 0.24, 0.33, 0.83, 1.41, 0.62, 0.28, 0.35, 0.77, 0.17, 0.72, 0.58, 0.45, 0.41), nrow=64,byrow=T) - triMutability_Names <- c("AAA", "AAC", "AAG", "AAT", "ACA", "ACC", "ACG", "ACT", "AGA", "AGC", "AGG", "AGT", "ATA", "ATC", "ATG", "ATT", "CAA", "CAC", "CAG", "CAT", "CCA", "CCC", "CCG", "CCT", "CGA", "CGC", "CGG", "CGT", "CTA", "CTC", "CTG", "CTT", "GAA", "GAC", "GAG", "GAT", "GCA", "GCC", "GCG", "GCT", "GGA", "GGC", "GGG", "GGT", "GTA", "GTC", "GTG", "GTT", "TAA", "TAC", "TAG", "TAT", "TCA", "TCC", "TCG", "TCT", "TGA", "TGC", "TGG", "TGT", "TTA", "TTC", "TTG", "TTT") - load("FiveS_Mutability.RData") - -# Functions - - # Translate codon to amino acid - translateCodonToAminoAcid<-function(Codon){ - return(AMINO_ACIDS[Codon]) - } - - # Translate amino acid to trait change - translateAminoAcidToTraitChange<-function(AminoAcid){ - return(TRAITS_AMINO_ACIDS[AminoAcid]) - } - - # Initialize Amino Acid Trait Changes - initializeTraitChange <- function(traitChangeModel=1,species=1,traitChangeFileName=NULL){ - if(!is.null(traitChangeFileName)){ - tryCatch( - traitChange <- read.delim(traitChangeFileName,sep="\t",header=T) - , error = function(ex){ - cat("Error|Error reading trait changes. Please check file name/path and format.\n") - q() - } - ) - }else{ - traitChange <- TRAITS_AMINO_ACIDS_CHOTHIA98 - } - TRAITS_AMINO_ACIDS <<- traitChange - } - - # Read in formatted nucleotide substitution matrix - initializeSubstitutionMatrix <- function(substitutionModel,species,subsMatFileName=NULL){ - if(!is.null(subsMatFileName)){ - tryCatch( - subsMat <- read.delim(subsMatFileName,sep="\t",header=T) - , error = function(ex){ - cat("Error|Error reading substitution matrix. Please check file name/path and format.\n") - q() - } - ) - if(sum(apply(subsMat,1,sum)==1)!=4) subsMat = t(apply(subsMat,1,function(x)x/sum(x))) - }else{ - if(substitutionModel==1)subsMat <- substitution_Literature_Mouse - if(substitutionModel==2)subsMat <- substitution_Flu_Human - if(substitutionModel==3)subsMat <- substitution_Flu25_Human - - } - - if(substitutionModel==0){ - subsMat <- matrix(1,4,4) - subsMat[,] = 1/3 - subsMat[1,1] = 0 - subsMat[2,2] = 0 - subsMat[3,3] = 0 - subsMat[4,4] = 0 - } - - - NUCLEOTIDESN = c(NUCLEOTIDES,"N", "-") - if(substitutionModel==5){ - subsMat <- FiveS_Substitution - return(subsMat) - }else{ - subsMat <- rbind(subsMat,rep(NA,4),rep(NA,4)) - return( matrix(data.matrix(subsMat),6,4,dimnames=list(NUCLEOTIDESN,NUCLEOTIDES) ) ) - } - } - - - # Read in formatted Mutability file - initializeMutabilityMatrix <- function(mutabilityModel=1, species=1,mutabilityMatFileName=NULL){ - if(!is.null(mutabilityMatFileName)){ - tryCatch( - mutabilityMat <- read.delim(mutabilityMatFileName,sep="\t",header=T) - , error = function(ex){ - cat("Error|Error reading mutability matrix. Please check file name/path and format.\n") - q() - } - ) - }else{ - mutabilityMat <- triMutability_Literature_Human - if(species==2) mutabilityMat <- triMutability_Literature_Mouse - } - - if(mutabilityModel==0){ mutabilityMat <- matrix(1,64,3)} - - if(mutabilityModel==5){ - mutabilityMat <- FiveS_Mutability - return(mutabilityMat) - }else{ - return( matrix( data.matrix(mutabilityMat), 64, 3, dimnames=list(triMutability_Names,1:3)) ) - } - } - - # Read FASTA file formats - # Modified from read.fasta from the seqinR package - baseline.read.fasta <- - function (file = system.file("sequences/sample.fasta", package = "seqinr"), - seqtype = c("DNA", "AA"), as.string = FALSE, forceDNAtolower = TRUE, - set.attributes = TRUE, legacy.mode = TRUE, seqonly = FALSE, - strip.desc = FALSE, sizeof.longlong = .Machine$sizeof.longlong, - endian = .Platform$endian, apply.mask = TRUE) - { - seqtype <- match.arg(seqtype) - - lines <- readLines(file) - - if (legacy.mode) { - comments <- grep("^;", lines) - if (length(comments) > 0) - lines <- lines[-comments] - } - - - ind_groups<-which(substr(lines, 1L, 3L) == ">>>") - lines_mod<-lines - - if(!length(ind_groups)){ - lines_mod<-c(">>>All sequences combined",lines) - } - - ind_groups<-which(substr(lines_mod, 1L, 3L) == ">>>") - - lines <- array("BLA",dim=(length(ind_groups)+length(lines_mod))) - id<-sapply(1:length(ind_groups),function(i)ind_groups[i]+i-1)+1 - lines[id] <- "THIS IS A FAKE SEQUENCE" - lines[-id] <- lines_mod - rm(lines_mod) - - ind <- which(substr(lines, 1L, 1L) == ">") - nseq <- length(ind) - if (nseq == 0) { - stop("no line starting with a > character found") - } - start <- ind + 1 - end <- ind - 1 - - while( any(which(ind%in%end)) ){ - ind=ind[-which(ind%in%end)] - nseq <- length(ind) - if (nseq == 0) { - stop("no line starting with a > character found") - } - start <- ind + 1 - end <- ind - 1 - } - - end <- c(end[-1], length(lines)) - sequences <- lapply(seq_len(nseq), function(i) paste(lines[start[i]:end[i]], collapse = "")) - if (seqonly) - return(sequences) - nomseq <- lapply(seq_len(nseq), function(i) { - - #firstword <- strsplit(lines[ind[i]], " ")[[1]][1] - substr(lines[ind[i]], 2, nchar(lines[ind[i]])) - - }) - if (seqtype == "DNA") { - if (forceDNAtolower) { - sequences <- as.list(tolower(chartr(".","-",sequences))) - }else{ - sequences <- as.list(toupper(chartr(".","-",sequences))) - } - } - if (as.string == FALSE) - sequences <- lapply(sequences, s2c) - if (set.attributes) { - for (i in seq_len(nseq)) { - Annot <- lines[ind[i]] - if (strip.desc) - Annot <- substr(Annot, 2L, nchar(Annot)) - attributes(sequences[[i]]) <- list(name = nomseq[[i]], - Annot = Annot, class = switch(seqtype, AA = "SeqFastaAA", - DNA = "SeqFastadna")) - } - } - names(sequences) <- nomseq - return(sequences) - } - - - # Replaces non FASTA characters in input files with N - replaceNonFASTAChars <-function(inSeq="ACGTN-AApA"){ - gsub('[^ACGTNacgt[:punct:]-[:punct:].]','N',inSeq,perl=TRUE) - } - - # Find the germlines in the FASTA list - germlinesInFile <- function(seqIDs){ - firstChar = sapply(seqIDs,function(x){substr(x,1,1)}) - secondChar = sapply(seqIDs,function(x){substr(x,2,2)}) - return(firstChar==">" & secondChar!=">") - } - - # Find the groups in the FASTA list - groupsInFile <- function(seqIDs){ - sapply(seqIDs,function(x){substr(x,1,2)})==">>" - } - - # In the process of finding germlines/groups, expand from the start to end of the group - expandTillNext <- function(vecPosToID){ - IDs = names(vecPosToID) - posOfInterests = which(vecPosToID) - - expandedID = rep(NA,length(IDs)) - expandedIDNames = gsub(">","",IDs[posOfInterests]) - startIndexes = c(1,posOfInterests[-1]) - stopIndexes = c(posOfInterests[-1]-1,length(IDs)) - expandedID = unlist(sapply(1:length(startIndexes),function(i){ - rep(i,stopIndexes[i]-startIndexes[i]+1) - })) - names(expandedID) = unlist(sapply(1:length(startIndexes),function(i){ - rep(expandedIDNames[i],stopIndexes[i]-startIndexes[i]+1) - })) - return(expandedID) - } - - # Process FASTA (list) to return a matrix[input, germline) - processInputAdvanced <- function(inputFASTA){ - - seqIDs = names(inputFASTA) - numbSeqs = length(seqIDs) - posGermlines1 = germlinesInFile(seqIDs) - numbGermlines = sum(posGermlines1) - posGroups1 = groupsInFile(seqIDs) - numbGroups = sum(posGroups1) - consDef = NA - - if(numbGermlines==0){ - posGermlines = 2 - numbGermlines = 1 - } - - glPositionsSum = cumsum(posGermlines1) - glPositions = table(glPositionsSum) - #Find the position of the conservation row - consDefPos = as.numeric(names(glPositions[names(glPositions)!=0 & glPositions==1]))+1 - if( length(consDefPos)> 0 ){ - consDefID = match(consDefPos, glPositionsSum) - #The coservation rows need to be pulled out and stores seperately - consDef = inputFASTA[consDefID] - inputFASTA = inputFASTA[-consDefID] - - seqIDs = names(inputFASTA) - numbSeqs = length(seqIDs) - posGermlines1 = germlinesInFile(seqIDs) - numbGermlines = sum(posGermlines1) - posGroups1 = groupsInFile(seqIDs) - numbGroups = sum(posGroups1) - if(numbGermlines==0){ - posGermlines = 2 - numbGermlines = 1 - } - } - - posGroups <- expandTillNext(posGroups1) - posGermlines <- expandTillNext(posGermlines1) - posGermlines[posGroups1] = 0 - names(posGermlines)[posGroups1] = names(posGroups)[posGroups1] - posInput = rep(TRUE,numbSeqs) - posInput[posGroups1 | posGermlines1] = FALSE - - matInput = matrix(NA, nrow=sum(posInput), ncol=2) - rownames(matInput) = seqIDs[posInput] - colnames(matInput) = c("Input","Germline") - - vecInputFASTA = unlist(inputFASTA) - matInput[,1] = vecInputFASTA[posInput] - matInput[,2] = vecInputFASTA[ which( names(inputFASTA)%in%paste(">",names(posGermlines)[posInput],sep="") )[ posGermlines[posInput]] ] - - germlines = posGermlines[posInput] - groups = posGroups[posInput] - - return( list("matInput"=matInput, "germlines"=germlines, "groups"=groups, "conservationDefinition"=consDef )) - } - - - # Replace leading and trailing dashes in the sequence - replaceLeadingTrailingDashes <- function(x,readEnd){ - iiGap = unlist(gregexpr("-",x[1])) - ggGap = unlist(gregexpr("-",x[2])) - #posToChange = intersect(iiGap,ggGap) - - - seqIn = replaceLeadingTrailingDashesHelper(x[1]) - seqGL = replaceLeadingTrailingDashesHelper(x[2]) - seqTemplate = rep('N',readEnd) - seqIn <- c(seqIn,seqTemplate[(length(seqIn)+1):readEnd]) - seqGL <- c(seqGL,seqTemplate[(length(seqGL)+1):readEnd]) -# if(posToChange!=-1){ -# seqIn[posToChange] = "-" -# seqGL[posToChange] = "-" -# } - - seqIn = c2s(seqIn[1:readEnd]) - seqGL = c2s(seqGL[1:readEnd]) - - lenGL = nchar(seqGL) - if(lenGL seqLen ) - trimmedSeq = substr(seqToTrim,1, ( (getCodonPos(seqLen)[1])-1 ) ) - - return(trimmedSeq) - } - - # Given a nuclotide position, returns the pos of the 3 nucs that made the codon - # e.g. nuc 86 is part of nucs 85,86,87 - getCodonPos <- function(nucPos){ - codonNum = (ceiling(nucPos/3))*3 - return( (codonNum-2):codonNum) - } - - # Given a nuclotide position, returns the codon number - # e.g. nuc 86 = codon 29 - getCodonNumb <- function(nucPos){ - return( ceiling(nucPos/3) ) - } - - # Given a codon, returns all the nuc positions that make the codon - getCodonNucs <- function(codonNumb){ - getCodonPos(codonNumb*3) - } - - computeCodonTable <- function(testID=1){ - - if(testID<=4){ - # Pre-compute every codons - intCounter = 1 - for(pOne in NUCLEOTIDES){ - for(pTwo in NUCLEOTIDES){ - for(pThree in NUCLEOTIDES){ - codon = paste(pOne,pTwo,pThree,sep="") - colnames(CODON_TABLE)[intCounter] = codon - intCounter = intCounter + 1 - CODON_TABLE[,codon] = mutationTypeOptimized(cbind(permutateAllCodon(codon),rep(codon,12))) - } - } - } - chars = c("N","A","C","G","T", "-") - for(a in chars){ - for(b in chars){ - for(c in chars){ - if(a=="N" | b=="N" | c=="N"){ - #cat(paste(a,b,c),sep="","\n") - CODON_TABLE[,paste(a,b,c,sep="")] = rep(NA,12) - } - } - } - } - - chars = c("-","A","C","G","T") - for(a in chars){ - for(b in chars){ - for(c in chars){ - if(a=="-" | b=="-" | c=="-"){ - #cat(paste(a,b,c),sep="","\n") - CODON_TABLE[,paste(a,b,c,sep="")] = rep(NA,12) - } - } - } - } - CODON_TABLE <<- as.matrix(CODON_TABLE) - } - } - - collapseClone <- function(vecInputSeqs,glSeq,readEnd,nonTerminalOnly=0){ - #print(length(vecInputSeqs)) - vecInputSeqs = unique(vecInputSeqs) - if(length(vecInputSeqs)==1){ - return( list( c(vecInputSeqs,glSeq), F) ) - }else{ - charInputSeqs <- sapply(vecInputSeqs, function(x){ - s2c(x)[1:readEnd] - }) - charGLSeq <- s2c(glSeq) - matClone <- sapply(1:readEnd, function(i){ - posNucs = unique(charInputSeqs[i,]) - posGL = charGLSeq[i] - error = FALSE - if(posGL=="-" & sum(!(posNucs%in%c("-","N")))==0 ){ - return(c("-",error)) - } - if(length(posNucs)==1) - return(c(posNucs[1],error)) - else{ - if("N"%in%posNucs){ - error=TRUE - } - if(sum(!posNucs[posNucs!="N"]%in%posGL)==0){ - return( c(posGL,error) ) - }else{ - #return( c(sample(posNucs[posNucs!="N"],1),error) ) - if(nonTerminalOnly==0){ - return( c(sample(charInputSeqs[i,charInputSeqs[i,]!="N" & charInputSeqs[i,]!=posGL],1),error) ) - }else{ - posNucs = charInputSeqs[i,charInputSeqs[i,]!="N" & charInputSeqs[i,]!=posGL] - posNucsTable = table(posNucs) - if(sum(posNucsTable>1)==0){ - return( c(posGL,error) ) - }else{ - return( c(sample( posNucs[posNucs%in%names(posNucsTable)[posNucsTable>1]],1),error) ) - } - } - - } - } - }) - - - #print(length(vecInputSeqs)) - return(list(c(c2s(matClone[1,]),glSeq),"TRUE"%in%matClone[2,])) - } - } - - # Compute the expected for each sequence-germline pair - getExpectedIndividual <- function(matInput){ - if( any(grep("multicore",search())) ){ - facGL <- factor(matInput[,2]) - facLevels = levels(facGL) - LisGLs_MutabilityU = mclapply(1:length(facLevels), function(x){ - computeMutabilities(facLevels[x]) - }) - facIndex = match(facGL,facLevels) - - LisGLs_Mutability = mclapply(1:nrow(matInput), function(x){ - cInput = rep(NA,nchar(matInput[x,1])) - cInput[s2c(matInput[x,1])!="N"] = 1 - LisGLs_MutabilityU[[facIndex[x]]] * cInput - }) - - LisGLs_Targeting = mclapply(1:dim(matInput)[1], function(x){ - computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) - }) - - LisGLs_MutationTypes = mclapply(1:length(matInput[,2]),function(x){ - #print(x) - computeMutationTypes(matInput[x,2]) - }) - - LisGLs_Exp = mclapply(1:dim(matInput)[1], function(x){ - computeExpected(LisGLs_Targeting[[x]],LisGLs_MutationTypes[[x]]) - }) - - ul_LisGLs_Exp = unlist(LisGLs_Exp) - return(matrix(ul_LisGLs_Exp,ncol=4,nrow=(length(ul_LisGLs_Exp)/4),byrow=T)) - }else{ - facGL <- factor(matInput[,2]) - facLevels = levels(facGL) - LisGLs_MutabilityU = lapply(1:length(facLevels), function(x){ - computeMutabilities(facLevels[x]) - }) - facIndex = match(facGL,facLevels) - - LisGLs_Mutability = lapply(1:nrow(matInput), function(x){ - cInput = rep(NA,nchar(matInput[x,1])) - cInput[s2c(matInput[x,1])!="N"] = 1 - LisGLs_MutabilityU[[facIndex[x]]] * cInput - }) - - LisGLs_Targeting = lapply(1:dim(matInput)[1], function(x){ - computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) - }) - - LisGLs_MutationTypes = lapply(1:length(matInput[,2]),function(x){ - #print(x) - computeMutationTypes(matInput[x,2]) - }) - - LisGLs_Exp = lapply(1:dim(matInput)[1], function(x){ - computeExpected(LisGLs_Targeting[[x]],LisGLs_MutationTypes[[x]]) - }) - - ul_LisGLs_Exp = unlist(LisGLs_Exp) - return(matrix(ul_LisGLs_Exp,ncol=4,nrow=(length(ul_LisGLs_Exp)/4),byrow=T)) - - } - } - - # Compute mutabilities of sequence based on the tri-nucleotide model - computeMutabilities <- function(paramSeq){ - seqLen = nchar(paramSeq) - seqMutabilites = rep(NA,seqLen) - - gaplessSeq = gsub("-", "", paramSeq) - gaplessSeqLen = nchar(gaplessSeq) - gaplessSeqMutabilites = rep(NA,gaplessSeqLen) - - if(mutabilityModel!=5){ - pos<- 3:(gaplessSeqLen) - subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) - gaplessSeqMutabilites[pos] = - tapply( c( - getMutability( substr(subSeq,1,3), 3) , - getMutability( substr(subSeq,2,4), 2), - getMutability( substr(subSeq,3,5), 1) - ),rep(1:(gaplessSeqLen-2),3),mean,na.rm=TRUE - ) - #Pos 1 - subSeq = substr(gaplessSeq,1,3) - gaplessSeqMutabilites[1] = getMutability(subSeq , 1) - #Pos 2 - subSeq = substr(gaplessSeq,1,4) - gaplessSeqMutabilites[2] = mean( c( - getMutability( substr(subSeq,1,3), 2) , - getMutability( substr(subSeq,2,4), 1) - ),na.rm=T - ) - seqMutabilites[which(s2c(paramSeq)!="-")]<- gaplessSeqMutabilites - return(seqMutabilites) - }else{ - - pos<- 3:(gaplessSeqLen) - subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) - gaplessSeqMutabilites[pos] = sapply(subSeq,function(x){ getMutability5(x) }, simplify=T) - seqMutabilites[which(s2c(paramSeq)!="-")]<- gaplessSeqMutabilites - return(seqMutabilites) - } - - } - - # Returns the mutability of a triplet at a given position - getMutability <- function(codon, pos=1:3){ - triplets <- rownames(mutability) - mutability[ match(codon,triplets) ,pos] - } - - getMutability5 <- function(fivemer){ - return(mutability[fivemer]) - } - - # Returns the substitution probabilty - getTransistionProb <- function(nuc){ - substitution[nuc,] - } - - getTransistionProb5 <- function(fivemer){ - if(any(which(fivemer==colnames(substitution)))){ - return(substitution[,fivemer]) - }else{ - return(array(NA,4)) - } - } - - # Given a nuc, returns the other 3 nucs it can mutate to - canMutateTo <- function(nuc){ - NUCLEOTIDES[- which(NUCLEOTIDES==nuc)] - } - - # Given a nucleotide, returns the probabilty of other nucleotide it can mutate to - canMutateToProb <- function(nuc){ - substitution[nuc,canMutateTo(nuc)] - } - - # Compute targeting, based on precomputed mutatbility & substitution - computeTargeting <- function(param_strSeq,param_vecMutabilities){ - - if(substitutionModel!=5){ - vecSeq = s2c(param_strSeq) - matTargeting = sapply( 1:length(vecSeq), function(x) { param_vecMutabilities[x] * getTransistionProb(vecSeq[x]) } ) - #matTargeting = apply( rbind(vecSeq,param_vecMutabilities),2, function(x) { as.vector(as.numeric(x[2]) * getTransistionProb(x[1])) } ) - dimnames( matTargeting ) = list(NUCLEOTIDES,1:(length(vecSeq))) - return (matTargeting) - }else{ - - seqLen = nchar(param_strSeq) - seqsubstitution = matrix(NA,ncol=seqLen,nrow=4) - paramSeq <- param_strSeq - gaplessSeq = gsub("-", "", paramSeq) - gaplessSeqLen = nchar(gaplessSeq) - gaplessSeqSubstitution = matrix(NA,ncol=gaplessSeqLen,nrow=4) - - pos<- 3:(gaplessSeqLen) - subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) - gaplessSeqSubstitution[,pos] = sapply(subSeq,function(x){ getTransistionProb5(x) }, simplify=T) - seqsubstitution[,which(s2c(paramSeq)!="-")]<- gaplessSeqSubstitution - #matTargeting <- param_vecMutabilities %*% seqsubstitution - matTargeting <- sweep(seqsubstitution,2,param_vecMutabilities,`*`) - dimnames( matTargeting ) = list(NUCLEOTIDES,1:(seqLen)) - return (matTargeting) - } - } - - # Compute the mutations types - computeMutationTypes <- function(param_strSeq){ - #cat(param_strSeq,"\n") - #vecSeq = trimToLastCodon(param_strSeq) - lenSeq = nchar(param_strSeq) - vecCodons = sapply({1:(lenSeq/3)}*3-2,function(x){substr(param_strSeq,x,x+2)}) - matMutationTypes = matrix( unlist(CODON_TABLE[,vecCodons]) ,ncol=lenSeq,nrow=4, byrow=F) - dimnames( matMutationTypes ) = list(NUCLEOTIDES,1:(ncol(matMutationTypes))) - return(matMutationTypes) - } - computeMutationTypesFast <- function(param_strSeq){ - matMutationTypes = matrix( CODON_TABLE[,param_strSeq] ,ncol=3,nrow=4, byrow=F) - #dimnames( matMutationTypes ) = list(NUCLEOTIDES,1:(length(vecSeq))) - return(matMutationTypes) - } - mutationTypeOptimized <- function( matOfCodons ){ - apply( matOfCodons,1,function(x){ mutationType(x[2],x[1]) } ) - } - - # Returns a vector of codons 1 mutation away from the given codon - permutateAllCodon <- function(codon){ - cCodon = s2c(codon) - matCodons = t(array(cCodon,dim=c(3,12))) - matCodons[1:4,1] = NUCLEOTIDES - matCodons[5:8,2] = NUCLEOTIDES - matCodons[9:12,3] = NUCLEOTIDES - apply(matCodons,1,c2s) - } - - # Given two codons, tells you if the mutation is R or S (based on your definition) - mutationType <- function(codonFrom,codonTo){ - if(testID==4){ - if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ - return(NA) - }else{ - mutationType = "S" - if( translateAminoAcidToTraitChange(translateCodonToAminoAcid(codonFrom)) != translateAminoAcidToTraitChange(translateCodonToAminoAcid(codonTo)) ){ - mutationType = "R" - } - if(translateCodonToAminoAcid(codonTo)=="*" | translateCodonToAminoAcid(codonFrom)=="*"){ - mutationType = "Stop" - } - return(mutationType) - } - }else if(testID==5){ - if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ - return(NA) - }else{ - if(codonFrom==codonTo){ - mutationType = "S" - }else{ - codonFrom = s2c(codonFrom) - codonTo = s2c(codonTo) - mutationType = "Stop" - nucOfI = codonFrom[which(codonTo!=codonFrom)] - if(nucOfI=="C"){ - mutationType = "R" - }else if(nucOfI=="G"){ - mutationType = "S" - } - } - return(mutationType) - } - }else{ - if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ - return(NA) - }else{ - mutationType = "S" - if( translateCodonToAminoAcid(codonFrom) != translateCodonToAminoAcid(codonTo) ){ - mutationType = "R" - } - if(translateCodonToAminoAcid(codonTo)=="*" | translateCodonToAminoAcid(codonFrom)=="*"){ - mutationType = "Stop" - } - return(mutationType) - } - } - } - - - #given a mat of targeting & it's corresponding mutationtypes returns - #a vector of Exp_RCDR,Exp_SCDR,Exp_RFWR,Exp_RFWR - computeExpected <- function(paramTargeting,paramMutationTypes){ - # Replacements - RPos = which(paramMutationTypes=="R") - #FWR - Exp_R_FWR = sum(paramTargeting[ RPos[which(FWR_Nuc_Mat[RPos]==T)] ],na.rm=T) - #CDR - Exp_R_CDR = sum(paramTargeting[ RPos[which(CDR_Nuc_Mat[RPos]==T)] ],na.rm=T) - # Silents - SPos = which(paramMutationTypes=="S") - #FWR - Exp_S_FWR = sum(paramTargeting[ SPos[which(FWR_Nuc_Mat[SPos]==T)] ],na.rm=T) - #CDR - Exp_S_CDR = sum(paramTargeting[ SPos[which(CDR_Nuc_Mat[SPos]==T)] ],na.rm=T) - - return(c(Exp_R_CDR,Exp_S_CDR,Exp_R_FWR,Exp_S_FWR)) - } - - # Count the mutations in a sequence - # each mutation is treated independently - analyzeMutations2NucUri_website <- function( rev_in_matrix ){ - paramGL = rev_in_matrix[2,] - paramSeq = rev_in_matrix[1,] - - #Fill seq with GL seq if gapped - #if( any(paramSeq=="-") ){ - # gapPos_Seq = which(paramSeq=="-") - # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "-"] - # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] - #} - - - #if( any(paramSeq=="N") ){ - # gapPos_Seq = which(paramSeq=="N") - # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] - # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] - #} - - analyzeMutations2NucUri( matrix(c( paramGL, paramSeq ),2,length(paramGL),byrow=T) ) - - } - - #1 = GL - #2 = Seq - analyzeMutations2NucUri <- function( in_matrix=matrix(c(c("A","A","A","C","C","C"),c("A","G","G","C","C","A")),2,6,byrow=T) ){ - paramGL = in_matrix[2,] - paramSeq = in_matrix[1,] - paramSeqUri = paramGL - #mutations = apply(rbind(paramGL,paramSeq), 2, function(x){!x[1]==x[2]}) - mutations_val = paramGL != paramSeq - if(any(mutations_val)){ - mutationPos = {1:length(mutations_val)}[mutations_val] - mutationPos = mutationPos[sapply(mutationPos, function(x){!any(paramSeq[getCodonPos(x)]=="N")})] - length_mutations =length(mutationPos) - mutationInfo = rep(NA,length_mutations) - if(any(mutationPos)){ - - pos<- mutationPos - pos_array<-array(sapply(pos,getCodonPos)) - codonGL = paramGL[pos_array] - - codonSeq = sapply(pos,function(x){ - seqP = paramGL[getCodonPos(x)] - muCodonPos = {x-1}%%3+1 - seqP[muCodonPos] = paramSeq[x] - return(seqP) - }) - GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) - Seqcodons = apply(codonSeq,2,c2s) - mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - names(mutationInfo) = mutationPos - } - if(any(!is.na(mutationInfo))){ - return(mutationInfo[!is.na(mutationInfo)]) - }else{ - return(NA) - } - - - }else{ - return (NA) - } - } - - processNucMutations2 <- function(mu){ - if(!is.na(mu)){ - #R - if(any(mu=="R")){ - Rs = mu[mu=="R"] - nucNumbs = as.numeric(names(Rs)) - R_CDR = sum(as.integer(CDR_Nuc[nucNumbs]),na.rm=T) - R_FWR = sum(as.integer(FWR_Nuc[nucNumbs]),na.rm=T) - }else{ - R_CDR = 0 - R_FWR = 0 - } - - #S - if(any(mu=="S")){ - Ss = mu[mu=="S"] - nucNumbs = as.numeric(names(Ss)) - S_CDR = sum(as.integer(CDR_Nuc[nucNumbs]),na.rm=T) - S_FWR = sum(as.integer(FWR_Nuc[nucNumbs]),na.rm=T) - }else{ - S_CDR = 0 - S_FWR = 0 - } - - - retVec = c(R_CDR,S_CDR,R_FWR,S_FWR) - retVec[is.na(retVec)]=0 - return(retVec) - }else{ - return(rep(0,4)) - } - } - - - ## Z-score Test - computeZScore <- function(mat, test="Focused"){ - matRes <- matrix(NA,ncol=2,nrow=(nrow(mat))) - if(test=="Focused"){ - #Z_Focused_CDR - #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) - P = apply(mat[,c(5,6,8)],1,function(x){(x[1]/sum(x))}) - R_mean = apply(cbind(mat[,c(1,2,4)],P),1,function(x){x[4]*(sum(x[1:3]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,1] = (mat[,1]-R_mean)/R_sd - - #Z_Focused_FWR - #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) - P = apply(mat[,c(7,6,8)],1,function(x){(x[1]/sum(x))}) - R_mean = apply(cbind(mat[,c(3,2,4)],P),1,function(x){x[4]*(sum(x[1:3]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,2] = (mat[,3]-R_mean)/R_sd - } - - if(test=="Local"){ - #Z_Focused_CDR - #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) - P = apply(mat[,c(5,6)],1,function(x){(x[1]/sum(x))}) - R_mean = apply(cbind(mat[,c(1,2)],P),1,function(x){x[3]*(sum(x[1:2]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,1] = (mat[,1]-R_mean)/R_sd - - #Z_Focused_FWR - #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) - P = apply(mat[,c(7,8)],1,function(x){(x[1]/sum(x))}) - R_mean = apply(cbind(mat[,c(3,4)],P),1,function(x){x[3]*(sum(x[1:2]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,2] = (mat[,3]-R_mean)/R_sd - } - - if(test=="Imbalanced"){ - #Z_Focused_CDR - #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) - P = apply(mat[,5:8],1,function(x){((x[1]+x[2])/sum(x))}) - R_mean = apply(cbind(mat[,1:4],P),1,function(x){x[5]*(sum(x[1:4]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,1] = (mat[,1]-R_mean)/R_sd - - #Z_Focused_FWR - #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) - P = apply(mat[,5:8],1,function(x){((x[3]+x[4])/sum(x))}) - R_mean = apply(cbind(mat[,1:4],P),1,function(x){x[5]*(sum(x[1:4]))}) - R_sd=sqrt(R_mean*(1-P)) - matRes[,2] = (mat[,3]-R_mean)/R_sd - } - - matRes[is.nan(matRes)] = NA - return(matRes) - } - - # Return a p-value for a z-score - z2p <- function(z){ - p=NA - if( !is.nan(z) && !is.na(z)){ - if(z>0){ - p = (1 - pnorm(z,0,1)) - } else if(z<0){ - p = (-1 * pnorm(z,0,1)) - } else{ - p = 0.5 - } - }else{ - p = NA - } - return(p) - } - - - ## Bayesian Test - - # Fitted parameter for the bayesian framework -BAYESIAN_FITTED<-c(0.407277142798302, 0.554007336744485, 0.63777155771234, 0.693989162719009, 0.735450014674917, 0.767972534429806, 0.794557287143399, 0.816906816601605, 0.83606796225341, 0.852729446430296, 0.867370424541641, 0.880339760590323, 0.891900995024999, 0.902259181289864, 0.911577919359,0.919990301665853, 0.927606458124537, 0.934518806350661, 0.940805863754375, 0.946534836475715, 0.951763691199255, 0.95654428191308, 0.960920179487397, 0.964930893680829, 0.968611312149038, 0.971992459313836, 0.975102110004818, 0.977964943023096, 0.980603428208439, 0.983037660179428, 0.985285800977406, 0.987364285326685, 0.989288037855441, 0.991070478823525, 0.992723699729969, 0.994259575477392, 0.995687688867975, 0.997017365051493, 0.998257085153047, 0.999414558305388, 1.00049681357804, 1.00151036237481, 1.00246080204981, 1.00335370751909, 1.0041939329768, 1.0049859393417, 1.00573382091263, 1.00644127217376, 1.00711179729107, 1.00774845526417, 1.00835412715854, 1.00893143010366, 1.00948275846309, 1.01001030293661, 1.01051606798079, 1.01100188771288, 1.01146944044216, 1.01192026195449, 1.01235575766094, 1.01277721370986) - CONST_i <- sort(c(((2^(seq(-39,0,length.out=201)))/2)[1:200],(c(0:11,13:99)+0.5)/100,1-(2^(seq(-39,0,length.out=201)))/2)) - - # Given x, M & p, returns a pdf - calculate_bayes <- function ( x=3, N=10, p=0.33, - i=CONST_i, - max_sigma=20,length_sigma=4001 - ){ - if(!0%in%N){ - G <- max(length(x),length(N),length(p)) - x=array(x,dim=G) - N=array(N,dim=G) - p=array(p,dim=G) - sigma_s<-seq(-max_sigma,max_sigma,length.out=length_sigma) - sigma_1<-log({i/{1-i}}/{p/{1-p}}) - index<-min(N,60) - y<-dbeta(i,x+BAYESIAN_FITTED[index],N+BAYESIAN_FITTED[index]-x)*(1-p)*p*exp(sigma_1)/({1-p}^2+2*p*{1-p}*exp(sigma_1)+{p^2}*exp(2*sigma_1)) - if(!sum(is.na(y))){ - tmp<-approx(sigma_1,y,sigma_s)$y - tmp/sum(tmp)/{2*max_sigma/{length_sigma-1}} - }else{ - return(NA) - } - }else{ - return(NA) - } - } - # Given a mat of observed & expected, return a list of CDR & FWR pdf for selection - computeBayesianScore <- function(mat, test="Focused", max_sigma=20,length_sigma=4001){ - flagOneSeq = F - if(nrow(mat)==1){ - mat=rbind(mat,mat) - flagOneSeq = T - } - if(test=="Focused"){ - #CDR - P = c(apply(mat[,c(5,6,8)],1,function(x){(x[1]/sum(x))}),0.5) - N = c(apply(mat[,c(1,2,4)],1,function(x){(sum(x))}),0) - X = c(mat[,1],0) - bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesCDR = bayesCDR[-length(bayesCDR)] - - #FWR - P = c(apply(mat[,c(7,6,8)],1,function(x){(x[1]/sum(x))}),0.5) - N = c(apply(mat[,c(3,2,4)],1,function(x){(sum(x))}),0) - X = c(mat[,3],0) - bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesFWR = bayesFWR[-length(bayesFWR)] - } - - if(test=="Local"){ - #CDR - P = c(apply(mat[,c(5,6)],1,function(x){(x[1]/sum(x))}),0.5) - N = c(apply(mat[,c(1,2)],1,function(x){(sum(x))}),0) - X = c(mat[,1],0) - bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesCDR = bayesCDR[-length(bayesCDR)] - - #FWR - P = c(apply(mat[,c(7,8)],1,function(x){(x[1]/sum(x))}),0.5) - N = c(apply(mat[,c(3,4)],1,function(x){(sum(x))}),0) - X = c(mat[,3],0) - bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesFWR = bayesFWR[-length(bayesFWR)] - } - - if(test=="Imbalanced"){ - #CDR - P = c(apply(mat[,c(5:8)],1,function(x){((x[1]+x[2])/sum(x))}),0.5) - N = c(apply(mat[,c(1:4)],1,function(x){(sum(x))}),0) - X = c(apply(mat[,c(1:2)],1,function(x){(sum(x))}),0) - bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesCDR = bayesCDR[-length(bayesCDR)] - - #FWR - P = c(apply(mat[,c(5:8)],1,function(x){((x[3]+x[4])/sum(x))}),0.5) - N = c(apply(mat[,c(1:4)],1,function(x){(sum(x))}),0) - X = c(apply(mat[,c(3:4)],1,function(x){(sum(x))}),0) - bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesFWR = bayesFWR[-length(bayesFWR)] - } - - if(test=="ImbalancedSilent"){ - #CDR - P = c(apply(mat[,c(6,8)],1,function(x){((x[1])/sum(x))}),0.5) - N = c(apply(mat[,c(2,4)],1,function(x){(sum(x))}),0) - X = c(apply(mat[,c(2,4)],1,function(x){(x[1])}),0) - bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesCDR = bayesCDR[-length(bayesCDR)] - - #FWR - P = c(apply(mat[,c(6,8)],1,function(x){((x[2])/sum(x))}),0.5) - N = c(apply(mat[,c(2,4)],1,function(x){(sum(x))}),0) - X = c(apply(mat[,c(2,4)],1,function(x){(x[2])}),0) - bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) - bayesFWR = bayesFWR[-length(bayesFWR)] - } - - if(flagOneSeq==T){ - bayesCDR = bayesCDR[1] - bayesFWR = bayesFWR[1] - } - return( list("CDR"=bayesCDR, "FWR"=bayesFWR) ) - } - - ##Covolution - break2chunks<-function(G=1000){ - base<-2^round(log(sqrt(G),2),0) - return(c(rep(base,floor(G/base)-1),base+G-(floor(G/base)*base))) - } - - PowersOfTwo <- function(G=100){ - exponents <- array() - i = 0 - while(G > 0){ - i=i+1 - exponents[i] <- floor( log2(G) ) - G <- G-2^exponents[i] - } - return(exponents) - } - - convolutionPowersOfTwo <- function( cons, length_sigma=4001 ){ - G = ncol(cons) - if(G>1){ - for(gen in log(G,2):1){ - ll<-seq(from=2,to=2^gen,by=2) - sapply(ll,function(l){cons[,l/2]<<-weighted_conv(cons[,l],cons[,l-1],length_sigma=length_sigma)}) - } - } - return( cons[,1] ) - } - - convolutionPowersOfTwoByTwos <- function( cons, length_sigma=4001,G=1 ){ - if(length(ncol(cons))) G<-ncol(cons) - groups <- PowersOfTwo(G) - matG <- matrix(NA, ncol=length(groups), nrow=length(cons)/G ) - startIndex = 1 - for( i in 1:length(groups) ){ - stopIndex <- 2^groups[i] + startIndex - 1 - if(stopIndex!=startIndex){ - matG[,i] <- convolutionPowersOfTwo( cons[,startIndex:stopIndex], length_sigma=length_sigma ) - startIndex = stopIndex + 1 - } - else { - if(G>1) matG[,i] <- cons[,startIndex:stopIndex] - else matG[,i] <- cons - #startIndex = stopIndex + 1 - } - } - return( list( matG, groups ) ) - } - - weighted_conv<-function(x,y,w=1,m=100,length_sigma=4001){ - lx<-length(x) - ly<-length(y) - if({lx1){ - while( i1 & Length_Postrior<=Threshold){ - cons = matrix(unlist(listPosteriors),length(listPosteriors[[1]]),length(listPosteriors)) - listMatG <- convolutionPowersOfTwoByTwos(cons,length_sigma=length_sigma) - y<-calculate_bayesGHelper(listMatG,length_sigma=length_sigma) - return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) - }else if(Length_Postrior==1) return(listPosteriors[[1]]) - else if(Length_Postrior==0) return(NA) - else { - cons = matrix(unlist(listPosteriors),length(listPosteriors[[1]]),length(listPosteriors)) - y = fastConv(cons,max_sigma=max_sigma, length_sigma=length_sigma ) - return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) - } - } - - fastConv<-function(cons, max_sigma=20, length_sigma=4001){ - chunks<-break2chunks(G=ncol(cons)) - if(ncol(cons)==3) chunks<-2:1 - index_chunks_end <- cumsum(chunks) - index_chunks_start <- c(1,index_chunks_end[-length(index_chunks_end)]+1) - index_chunks <- cbind(index_chunks_start,index_chunks_end) - - case <- sum(chunks!=chunks[1]) - if(case==1) End <- max(1,((length(index_chunks)/2)-1)) - else End <- max(1,((length(index_chunks)/2))) - - firsts <- sapply(1:End,function(i){ - indexes<-index_chunks[i,1]:index_chunks[i,2] - convolutionPowersOfTwoByTwos(cons[ ,indexes])[[1]] - }) - if(case==0){ - result<-calculate_bayesGHelper( convolutionPowersOfTwoByTwos(firsts) ) - }else if(case==1){ - last<-list(calculate_bayesGHelper( - convolutionPowersOfTwoByTwos( cons[ ,index_chunks[length(index_chunks)/2,1]:index_chunks[length(index_chunks)/2,2]] ) - ),0) - result_first<-calculate_bayesGHelper(convolutionPowersOfTwoByTwos(firsts)) - result<-calculate_bayesGHelper( - list( - cbind( - result_first,last[[1]]), - c(log(index_chunks_end[length(index_chunks)/2-1],2),log(index_chunks[length(index_chunks)/2,2]-index_chunks[length(index_chunks)/2,1]+1,2)) - ) - ) - } - return(as.vector(result)) - } - - # Computes the 95% CI for a pdf - calcBayesCI <- function(Pdf,low=0.025,up=0.975,max_sigma=20, length_sigma=4001){ - if(length(Pdf)!=length_sigma) return(NA) - sigma_s=seq(-max_sigma,max_sigma,length.out=length_sigma) - cdf = cumsum(Pdf) - cdf = cdf/cdf[length(cdf)] - return( c(sigma_s[findInterval(low,cdf)-1] , sigma_s[findInterval(up,cdf)]) ) - } - - # Computes a mean for a pdf - calcBayesMean <- function(Pdf,max_sigma=20,length_sigma=4001){ - if(length(Pdf)!=length_sigma) return(NA) - sigma_s=seq(-max_sigma,max_sigma,length.out=length_sigma) - norm = {length_sigma-1}/2/max_sigma - return( (Pdf%*%sigma_s/norm) ) - } - - # Returns the mean, and the 95% CI for a pdf - calcBayesOutputInfo <- function(Pdf,low=0.025,up=0.975,max_sigma=20, length_sigma=4001){ - if(is.na(Pdf)) - return(rep(NA,3)) - bCI = calcBayesCI(Pdf=Pdf,low=low,up=up,max_sigma=max_sigma,length_sigma=length_sigma) - bMean = calcBayesMean(Pdf=Pdf,max_sigma=max_sigma,length_sigma=length_sigma) - return(c(bMean, bCI)) - } - - # Computes the p-value of a pdf - computeSigmaP <- function(Pdf, length_sigma=4001, max_sigma=20){ - if(length(Pdf)>1){ - norm = {length_sigma-1}/2/max_sigma - pVal = {sum(Pdf[1:{{length_sigma-1}/2}]) + Pdf[{{length_sigma+1}/2}]/2}/norm - if(pVal>0.5){ - pVal = pVal-1 - } - return(pVal) - }else{ - return(NA) - } - } - - # Compute p-value of two distributions - compareTwoDistsFaster <-function(sigma_S=seq(-20,20,length.out=4001), N=10000, dens1=runif(4001,0,1), dens2=runif(4001,0,1)){ - #print(c(length(dens1),length(dens2))) - if(length(dens1)>1 & length(dens2)>1 ){ - dens1<-dens1/sum(dens1) - dens2<-dens2/sum(dens2) - cum2 <- cumsum(dens2)-dens2/2 - tmp<- sum(sapply(1:length(dens1),function(i)return(dens1[i]*cum2[i]))) - #print(tmp) - if(tmp>0.5)tmp<-tmp-1 - return( tmp ) - } - else { - return(NA) - } - #return (sum(sapply(1:N,function(i)(sample(sigma_S,1,prob=dens1)>sample(sigma_S,1,prob=dens2))))/N) - } - - # get number of seqeunces contributing to the sigma (i.e. seqeunces with mutations) - numberOfSeqsWithMutations <- function(matMutations,test=1){ - if(test==4)test=2 - cdrSeqs <- 0 - fwrSeqs <- 0 - if(test==1){#focused - cdrMutations <- apply(matMutations, 1, function(x){ sum(x[c(1,2,4)]) }) - fwrMutations <- apply(matMutations, 1, function(x){ sum(x[c(3,4,2)]) }) - if( any(which(cdrMutations>0)) ) cdrSeqs <- sum(cdrMutations>0) - if( any(which(fwrMutations>0)) ) fwrSeqs <- sum(fwrMutations>0) - } - if(test==2){#local - cdrMutations <- apply(matMutations, 1, function(x){ sum(x[c(1,2)]) }) - fwrMutations <- apply(matMutations, 1, function(x){ sum(x[c(3,4)]) }) - if( any(which(cdrMutations>0)) ) cdrSeqs <- sum(cdrMutations>0) - if( any(which(fwrMutations>0)) ) fwrSeqs <- sum(fwrMutations>0) - } - return(c("CDR"=cdrSeqs, "FWR"=fwrSeqs)) -} - - - -shadeColor <- function(sigmaVal=NA,pVal=NA){ - if(is.na(sigmaVal) & is.na(pVal)) return(NA) - if(is.na(sigmaVal) & !is.na(pVal)) sigmaVal=sign(pVal) - if(is.na(pVal) || pVal==1 || pVal==0){ - returnColor = "#FFFFFF"; - }else{ - colVal=abs(pVal); - - if(sigmaVal<0){ - if(colVal>0.1) - returnColor = "#CCFFCC"; - if(colVal<=0.1) - returnColor = "#99FF99"; - if(colVal<=0.050) - returnColor = "#66FF66"; - if(colVal<=0.010) - returnColor = "#33FF33"; - if(colVal<=0.005) - returnColor = "#00FF00"; - - }else{ - if(colVal>0.1) - returnColor = "#FFCCCC"; - if(colVal<=0.1) - returnColor = "#FF9999"; - if(colVal<=0.05) - returnColor = "#FF6666"; - if(colVal<=0.01) - returnColor = "#FF3333"; - if(colVal<0.005) - returnColor = "#FF0000"; - } - } - - return(returnColor) -} - - - -plotHelp <- function(xfrac=0.05,yfrac=0.05,log=FALSE){ - if(!log){ - x = par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac - y = par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac - }else { - if(log==2){ - x = par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac - y = 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac) - } - if(log==1){ - x = 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac) - y = par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac - } - if(log==3){ - x = 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac) - y = 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac) - } - } - return(c("x"=x,"y"=y)) -} - -# SHMulation - - # Based on targeting, introduce a single mutation & then update the targeting - oneMutation <- function(){ - # Pick a postion + mutation - posMutation = sample(1:(seqGermlineLen*4),1,replace=F,prob=as.vector(seqTargeting)) - posNucNumb = ceiling(posMutation/4) # Nucleotide number - posNucKind = 4 - ( (posNucNumb*4) - posMutation ) # Nuc the position mutates to - - #mutate the simulation sequence - seqSimVec <- s2c(seqSim) - seqSimVec[posNucNumb] <- NUCLEOTIDES[posNucKind] - seqSim <<- c2s(seqSimVec) - - #update Mutability, Targeting & MutationsTypes - updateMutabilityNTargeting(posNucNumb) - - #return(c(posNucNumb,NUCLEOTIDES[posNucKind])) - return(posNucNumb) - } - - updateMutabilityNTargeting <- function(position){ - min_i<-max((position-2),1) - max_i<-min((position+2),nchar(seqSim)) - min_ii<-min(min_i,3) - - #mutability - update locally - seqMutability[(min_i):(max_i)] <<- computeMutabilities(substr(seqSim,position-4,position+4))[(min_ii):(max_i-min_i+min_ii)] - - - #targeting - compute locally - seqTargeting[,min_i:max_i] <<- computeTargeting(substr(seqSim,min_i,max_i),seqMutability[min_i:max_i]) - seqTargeting[is.na(seqTargeting)] <<- 0 - #mutCodonPos = getCodonPos(position) - mutCodonPos = seq(getCodonPos(min_i)[1],getCodonPos(max_i)[3]) - #cat(mutCodonPos,"\n") - mutTypeCodon = getCodonPos(position) - seqMutationTypes[,mutTypeCodon] <<- computeMutationTypesFast( substr(seqSim,mutTypeCodon[1],mutTypeCodon[3]) ) - # Stop = 0 - if(any(seqMutationTypes[,mutCodonPos]=="Stop",na.rm=T )){ - seqTargeting[,mutCodonPos][seqMutationTypes[,mutCodonPos]=="Stop"] <<- 0 - } - - - #Selection - selectedPos = (min_i*4-4)+(which(seqMutationTypes[,min_i:max_i]=="R")) - # CDR - selectedCDR = selectedPos[which(matCDR[selectedPos]==T)] - seqTargeting[selectedCDR] <<- seqTargeting[selectedCDR] * exp(selCDR) - seqTargeting[selectedCDR] <<- seqTargeting[selectedCDR]/baseLineCDR_K - - # FWR - selectedFWR = selectedPos[which(matFWR[selectedPos]==T)] - seqTargeting[selectedFWR] <<- seqTargeting[selectedFWR] * exp(selFWR) - seqTargeting[selectedFWR] <<- seqTargeting[selectedFWR]/baseLineFWR_K - - } - - - - # Validate the mutation: if the mutation has not been sampled before validate it, else discard it. - validateMutation <- function(){ - if( !(mutatedPos%in%mutatedPositions) ){ # if it's a new mutation - uniqueMutationsIntroduced <<- uniqueMutationsIntroduced + 1 - mutatedPositions[uniqueMutationsIntroduced] <<- mutatedPos - }else{ - if(substr(seqSim,mutatedPos,mutatedPos)==substr(seqGermline,mutatedPos,mutatedPos)){ # back to germline mutation - mutatedPositions <<- mutatedPositions[-which(mutatedPositions==mutatedPos)] - uniqueMutationsIntroduced <<- uniqueMutationsIntroduced - 1 - } - } - } - - - - # Places text (labels) at normalized coordinates - myaxis <- function(xfrac=0.05,yfrac=0.05,log=FALSE,w="text",cex=1,adj=1,thecol="black"){ - par(xpd=TRUE) - if(!log) - text(par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac,par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac,w,cex=cex,adj=adj,col=thecol) - else { - if(log==2) - text( - par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac, - 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac), - w,cex=cex,adj=adj,col=thecol) - if(log==1) - text( - 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac), - par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac, - w,cex=cex,adj=adj,col=thecol) - if(log==3) - text( - 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac), - 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac), - w,cex=cex,adj=adj,col=thecol) - } - par(xpd=FALSE) - } - - - - # Count the mutations in a sequence - analyzeMutations <- function( inputMatrixIndex, model = 0 , multipleMutation=0, seqWithStops=0){ - - paramGL = s2c(matInput[inputMatrixIndex,2]) - paramSeq = s2c(matInput[inputMatrixIndex,1]) - - #if( any(paramSeq=="N") ){ - # gapPos_Seq = which(paramSeq=="N") - # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] - # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] - #} - mutations_val = paramGL != paramSeq - - if(any(mutations_val)){ - mutationPos = which(mutations_val)#{1:length(mutations_val)}[mutations_val] - length_mutations =length(mutationPos) - mutationInfo = rep(NA,length_mutations) - - pos<- mutationPos - pos_array<-array(sapply(pos,getCodonPos)) - codonGL = paramGL[pos_array] - codonSeqWhole = paramSeq[pos_array] - codonSeq = sapply(pos,function(x){ - seqP = paramGL[getCodonPos(x)] - muCodonPos = {x-1}%%3+1 - seqP[muCodonPos] = paramSeq[x] - return(seqP) - }) - GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) - SeqcodonsWhole = apply(matrix(codonSeqWhole,length_mutations,3,byrow=TRUE),1,c2s) - Seqcodons = apply(codonSeq,2,c2s) - - mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - names(mutationInfo) = mutationPos - - mutationInfoWhole = apply(rbind(GLcodons , SeqcodonsWhole),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - names(mutationInfoWhole) = mutationPos - - mutationInfo <- mutationInfo[!is.na(mutationInfo)] - mutationInfoWhole <- mutationInfoWhole[!is.na(mutationInfoWhole)] - - if(any(!is.na(mutationInfo))){ - - #Filter based on Stop (at the codon level) - if(seqWithStops==1){ - nucleotidesAtStopCodons = names(mutationInfoWhole[mutationInfoWhole!="Stop"]) - mutationInfo = mutationInfo[nucleotidesAtStopCodons] - mutationInfoWhole = mutationInfo[nucleotidesAtStopCodons] - }else{ - countStops = sum(mutationInfoWhole=="Stop") - if(seqWithStops==2 & countStops==0) mutationInfo = NA - if(seqWithStops==3 & countStops>0) mutationInfo = NA - } - - if(any(!is.na(mutationInfo))){ - #Filter mutations based on multipleMutation - if(multipleMutation==1 & !is.na(mutationInfo)){ - mutationCodons = getCodonNumb(as.numeric(names(mutationInfoWhole))) - tableMutationCodons <- table(mutationCodons) - codonsWithMultipleMutations <- as.numeric(names(tableMutationCodons[tableMutationCodons>1])) - if(any(codonsWithMultipleMutations)){ - #remove the nucleotide mutations in the codons with multiple mutations - mutationInfo <- mutationInfo[!(mutationCodons %in% codonsWithMultipleMutations)] - #replace those codons with Ns in the input sequence - paramSeq[unlist(lapply(codonsWithMultipleMutations, getCodonNucs))] = "N" - matInput[inputMatrixIndex,1] <<- c2s(paramSeq) - } - } - - #Filter mutations based on the model - if(any(mutationInfo)==T | is.na(any(mutationInfo))){ - - if(model==1 & !is.na(mutationInfo)){ - mutationInfo <- mutationInfo[mutationInfo=="S"] - } - if(any(mutationInfo)==T | is.na(any(mutationInfo))) return(mutationInfo) - else return(NA) - }else{ - return(NA) - } - }else{ - return(NA) - } - - - }else{ - return(NA) - } - - - }else{ - return (NA) - } - } - - analyzeMutationsFixed <- function( inputArray, model = 0 , multipleMutation=0, seqWithStops=0){ - - paramGL = s2c(inputArray[2]) - paramSeq = s2c(inputArray[1]) - inputSeq <- inputArray[1] - #if( any(paramSeq=="N") ){ - # gapPos_Seq = which(paramSeq=="N") - # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] - # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] - #} - mutations_val = paramGL != paramSeq - - if(any(mutations_val)){ - mutationPos = which(mutations_val)#{1:length(mutations_val)}[mutations_val] - length_mutations =length(mutationPos) - mutationInfo = rep(NA,length_mutations) - - pos<- mutationPos - pos_array<-array(sapply(pos,getCodonPos)) - codonGL = paramGL[pos_array] - codonSeqWhole = paramSeq[pos_array] - codonSeq = sapply(pos,function(x){ - seqP = paramGL[getCodonPos(x)] - muCodonPos = {x-1}%%3+1 - seqP[muCodonPos] = paramSeq[x] - return(seqP) - }) - GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) - SeqcodonsWhole = apply(matrix(codonSeqWhole,length_mutations,3,byrow=TRUE),1,c2s) - Seqcodons = apply(codonSeq,2,c2s) - - mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - names(mutationInfo) = mutationPos - - mutationInfoWhole = apply(rbind(GLcodons , SeqcodonsWhole),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - names(mutationInfoWhole) = mutationPos - - mutationInfo <- mutationInfo[!is.na(mutationInfo)] - mutationInfoWhole <- mutationInfoWhole[!is.na(mutationInfoWhole)] - - if(any(!is.na(mutationInfo))){ - - #Filter based on Stop (at the codon level) - if(seqWithStops==1){ - nucleotidesAtStopCodons = names(mutationInfoWhole[mutationInfoWhole!="Stop"]) - mutationInfo = mutationInfo[nucleotidesAtStopCodons] - mutationInfoWhole = mutationInfo[nucleotidesAtStopCodons] - }else{ - countStops = sum(mutationInfoWhole=="Stop") - if(seqWithStops==2 & countStops==0) mutationInfo = NA - if(seqWithStops==3 & countStops>0) mutationInfo = NA - } - - if(any(!is.na(mutationInfo))){ - #Filter mutations based on multipleMutation - if(multipleMutation==1 & !is.na(mutationInfo)){ - mutationCodons = getCodonNumb(as.numeric(names(mutationInfoWhole))) - tableMutationCodons <- table(mutationCodons) - codonsWithMultipleMutations <- as.numeric(names(tableMutationCodons[tableMutationCodons>1])) - if(any(codonsWithMultipleMutations)){ - #remove the nucleotide mutations in the codons with multiple mutations - mutationInfo <- mutationInfo[!(mutationCodons %in% codonsWithMultipleMutations)] - #replace those codons with Ns in the input sequence - paramSeq[unlist(lapply(codonsWithMultipleMutations, getCodonNucs))] = "N" - #matInput[inputMatrixIndex,1] <<- c2s(paramSeq) - inputSeq <- c2s(paramSeq) - } - } - - #Filter mutations based on the model - if(any(mutationInfo)==T | is.na(any(mutationInfo))){ - - if(model==1 & !is.na(mutationInfo)){ - mutationInfo <- mutationInfo[mutationInfo=="S"] - } - if(any(mutationInfo)==T | is.na(any(mutationInfo))) return(list(mutationInfo,inputSeq)) - else return(list(NA,inputSeq)) - }else{ - return(list(NA,inputSeq)) - } - }else{ - return(list(NA,inputSeq)) - } - - - }else{ - return(list(NA,inputSeq)) - } - - - }else{ - return (list(NA,inputSeq)) - } - } - - # triMutability Background Count - buildMutabilityModel <- function( inputMatrixIndex, model=0 , multipleMutation=0, seqWithStops=0, stopMutations=0){ - - #rowOrigMatInput = matInput[inputMatrixIndex,] - seqGL = gsub("-", "", matInput[inputMatrixIndex,2]) - seqInput = gsub("-", "", matInput[inputMatrixIndex,1]) - #matInput[inputMatrixIndex,] <<- cbind(seqInput,seqGL) - tempInput <- cbind(seqInput,seqGL) - seqLength = nchar(seqGL) - list_analyzeMutationsFixed<- analyzeMutationsFixed(tempInput, model, multipleMutation, seqWithStops) - mutationCount <- list_analyzeMutationsFixed[[1]] - seqInput <- list_analyzeMutationsFixed[[2]] - BackgroundMatrix = mutabilityMatrix - MutationMatrix = mutabilityMatrix - MutationCountMatrix = mutabilityMatrix - if(!is.na(mutationCount)){ - if((stopMutations==0 & model==0) | (stopMutations==1 & (sum(mutationCount=="Stop")0)) ){ - - fivermerStartPos = 1:(seqLength-4) - fivemerLength <- length(fivermerStartPos) - fivemerGL <- substr(rep(seqGL,length(fivermerStartPos)),(fivermerStartPos),(fivermerStartPos+4)) - fivemerSeq <- substr(rep(seqInput,length(fivermerStartPos)),(fivermerStartPos),(fivermerStartPos+4)) - - #Background - for(fivemerIndex in 1:fivemerLength){ - fivemer = fivemerGL[fivemerIndex] - if(!any(grep("N",fivemer))){ - fivemerCodonPos = fivemerCodon(fivemerIndex) - fivemerReadingFrameCodon = substr(fivemer,fivemerCodonPos[1],fivemerCodonPos[3]) - fivemerReadingFrameCodonInputSeq = substr(fivemerSeq[fivemerIndex],fivemerCodonPos[1],fivemerCodonPos[3]) - - # All mutations model - #if(!any(grep("N",fivemerReadingFrameCodon))){ - if(model==0){ - if(stopMutations==0){ - if(!any(grep("N",fivemerReadingFrameCodonInputSeq))) - BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + 1) - }else{ - if( !any(grep("N",fivemerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(fivemerReadingFrameCodon)!="*" ){ - positionWithinCodon = which(fivemerCodonPos==3)#positionsWithinCodon[(fivemerCodonPos[1]%%3)+1] - BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + probNonStopMutations[fivemerReadingFrameCodon,positionWithinCodon]) - } - } - }else{ # Only silent mutations - if( !any(grep("N",fivemerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(fivemerReadingFrameCodon)!="*" & translateCodonToAminoAcid(fivemerReadingFrameCodonInputSeq)==translateCodonToAminoAcid(fivemerReadingFrameCodon)){ - positionWithinCodon = which(fivemerCodonPos==3) - BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + probSMutations[fivemerReadingFrameCodon,positionWithinCodon]) - } - } - #} - } - } - - #Mutations - if(stopMutations==1) mutationCount = mutationCount[mutationCount!="Stop"] - if(model==1) mutationCount = mutationCount[mutationCount=="S"] - mutationPositions = as.numeric(names(mutationCount)) - mutationCount = mutationCount[mutationPositions>2 & mutationPositions<(seqLength-1)] - mutationPositions = mutationPositions[mutationPositions>2 & mutationPositions<(seqLength-1)] - countMutations = 0 - for(mutationPosition in mutationPositions){ - fivemerIndex = mutationPosition-2 - fivemer = fivemerSeq[fivemerIndex] - GLfivemer = fivemerGL[fivemerIndex] - fivemerCodonPos = fivemerCodon(fivemerIndex) - fivemerReadingFrameCodon = substr(fivemer,fivemerCodonPos[1],fivemerCodonPos[3]) - fivemerReadingFrameCodonGL = substr(GLfivemer,fivemerCodonPos[1],fivemerCodonPos[3]) - if(!any(grep("N",fivemer)) & !any(grep("N",GLfivemer))){ - if(model==0){ - countMutations = countMutations + 1 - MutationMatrix[GLfivemer] <- (MutationMatrix[GLfivemer] + 1) - MutationCountMatrix[GLfivemer] <- (MutationCountMatrix[GLfivemer] + 1) - }else{ - if( translateCodonToAminoAcid(fivemerReadingFrameCodonGL)!="*" ){ - countMutations = countMutations + 1 - positionWithinCodon = which(fivemerCodonPos==3) - glNuc = substr(fivemerReadingFrameCodonGL,positionWithinCodon,positionWithinCodon) - inputNuc = substr(fivemerReadingFrameCodon,positionWithinCodon,positionWithinCodon) - MutationMatrix[GLfivemer] <- (MutationMatrix[GLfivemer] + substitution[glNuc,inputNuc]) - MutationCountMatrix[GLfivemer] <- (MutationCountMatrix[GLfivemer] + 1) - } - } - } - } - - seqMutability = MutationMatrix/BackgroundMatrix - seqMutability = seqMutability/sum(seqMutability,na.rm=TRUE) - #cat(inputMatrixIndex,"\t",countMutations,"\n") - return(list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix, "BackgroundMatrix"=BackgroundMatrix)) - - } - } - - } - - #Returns the codon position containing the middle nucleotide - fivemerCodon <- function(fivemerIndex){ - codonPos = list(2:4,1:3,3:5) - fivemerType = fivemerIndex%%3 - return(codonPos[[fivemerType+1]]) - } - - #returns probability values for one mutation in codons resulting in R, S or Stop - probMutations <- function(typeOfMutation){ - matMutationProb <- matrix(0,ncol=3,nrow=125,dimnames=list(words(alphabet = c(NUCLEOTIDES,"N"), length=3),c(1:3))) - for(codon in rownames(matMutationProb)){ - if( !any(grep("N",codon)) ){ - for(muPos in 1:3){ - matCodon = matrix(rep(s2c(codon),3),nrow=3,ncol=3,byrow=T) - glNuc = matCodon[1,muPos] - matCodon[,muPos] = canMutateTo(glNuc) - substitutionRate = substitution[glNuc,matCodon[,muPos]] - typeOfMutations = apply(rbind(rep(codon,3),apply(matCodon,1,c2s)),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) - matMutationProb[codon,muPos] <- sum(substitutionRate[typeOfMutations==typeOfMutation]) - } - } - } - - return(matMutationProb) - } - - - - -#Mapping Trinucleotides to fivemers -mapTriToFivemer <- function(triMutability=triMutability_Literature_Human){ - rownames(triMutability) <- triMutability_Names - Fivemer<-rep(NA,1024) - names(Fivemer)<-words(alphabet=NUCLEOTIDES,length=5) - Fivemer<-sapply(names(Fivemer),function(Word)return(sum( c(triMutability[substring(Word,3,5),1],triMutability[substring(Word,2,4),2],triMutability[substring(Word,1,3),3]),na.rm=TRUE))) - Fivemer<-Fivemer/sum(Fivemer) - return(Fivemer) -} - -collapseFivemerToTri<-function(Fivemer,Weights=MutabilityWeights,position=1,NUC="A"){ - Indices<-substring(names(Fivemer),3,3)==NUC - Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) - tapply(which(Indices),Factors,function(i)weighted.mean(Fivemer[i],Weights[i],na.rm=TRUE)) -} - - - -CountFivemerToTri<-function(Fivemer,Weights=MutabilityWeights,position=1,NUC="A"){ - Indices<-substring(names(Fivemer),3,3)==NUC - Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) - tapply(which(Indices),Factors,function(i)sum(Weights[i],na.rm=TRUE)) -} - -#Uses the real counts of the mutated fivemers -CountFivemerToTri2<-function(Fivemer,Counts=MutabilityCounts,position=1,NUC="A"){ - Indices<-substring(names(Fivemer),3,3)==NUC - Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) - tapply(which(Indices),Factors,function(i)sum(Counts[i],na.rm=TRUE)) -} - -bootstrap<-function(x=c(33,12,21),M=10000,alpha=0.05){ -N<-sum(x) -if(N){ -p<-x/N -k<-length(x)-1 -tmp<-rmultinom(M, size = N, prob=p) -tmp_p<-apply(tmp,2,function(y)y/N) -(apply(tmp_p,1,function(y)quantile(y,c(alpha/2/k,1-alpha/2/k)))) -} -else return(matrix(0,2,length(x))) -} - - - - -bootstrap2<-function(x=c(33,12,21),n=10,M=10000,alpha=0.05){ - -N<-sum(x) -k<-length(x) -y<-rep(1:k,x) -tmp<-sapply(1:M,function(i)sample(y,n)) -if(n>1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[,j]==i)))/n -if(n==1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[j]==i)))/n -(apply(tmp_p,1,function(z)quantile(z,c(alpha/2/(k-1),1-alpha/2/(k-1))))) -} - - - -p_value<-function(x=c(33,12,21),M=100000,x_obs=c(2,5,3)){ -n=sum(x_obs) -N<-sum(x) -k<-length(x) -y<-rep(1:k,x) -tmp<-sapply(1:M,function(i)sample(y,n)) -if(n>1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[,j]==i))) -if(n==1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[j]==i))) -tmp<-rbind(sapply(1:3,function(i)sum(tmp_p[i,]>=x_obs[i])/M), -sapply(1:3,function(i)sum(tmp_p[i,]<=x_obs[i])/M)) -sapply(1:3,function(i){if(tmp[1,i]>=tmp[2,i])return(-tmp[2,i])else return(tmp[1,i])}) -} - -#"D:\\Sequences\\IMGT Germlines\\Human_SNPless_IGHJ.FASTA" -# Remove SNPs from IMGT germline segment alleles -generateUnambiguousRepertoire <- function(repertoireInFile,repertoireOutFile){ - repertoireIn <- read.fasta(repertoireInFile, seqtype="DNA",as.string=T,set.attributes=F,forceDNAtolower=F) - alleleNames <- sapply(names(repertoireIn),function(x)strsplit(x,"|",fixed=TRUE)[[1]][2]) - SNPs <- tapply(repertoireIn,sapply(alleleNames,function(x)strsplit(x,"*",fixed=TRUE)[[1]][1]),function(x){ - Indices<-NULL - for(i in 1:length(x)){ - firstSeq = s2c(x[[1]]) - iSeq = s2c(x[[i]]) - Indices<-c(Indices,which(firstSeq[1:320]!=iSeq[1:320] & firstSeq[1:320]!="." & iSeq[1:320]!="." )) - } - return(sort(unique(Indices))) - }) - repertoireOut <- repertoireIn - repertoireOut <- lapply(names(repertoireOut), function(repertoireName){ - alleleName <- strsplit(repertoireName,"|",fixed=TRUE)[[1]][2] - geneSegmentName <- strsplit(alleleName,"*",fixed=TRUE)[[1]][1] - alleleSeq <- s2c(repertoireOut[[repertoireName]]) - alleleSeq[as.numeric(unlist(SNPs[geneSegmentName]))] <- "N" - alleleSeq <- c2s(alleleSeq) - repertoireOut[[repertoireName]] <- alleleSeq - }) - names(repertoireOut) <- names(repertoireIn) - write.fasta(repertoireOut,names(repertoireOut),file.out=repertoireOutFile) - -} - - - - - - -############ -groupBayes2 = function(indexes, param_resultMat){ - - BayesGDist_Focused_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2]+x[4])})) - BayesGDist_Focused_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(3,2,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[3]+x[2]+x[4])})) - #BayesGDist_Local_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2])})) - #BayesGDist_Local_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[3]+x[4])})) - #BayesGDist_Global_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2,3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2]+x[3]+x[4])})) - #BayesGDist_Global_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(1,2,3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[1]+x[2]+x[3]+x[4])})) - return ( list("BayesGDist_Focused_CDR"=BayesGDist_Focused_CDR, - "BayesGDist_Focused_FWR"=BayesGDist_Focused_FWR) ) - #"BayesGDist_Local_CDR"=BayesGDist_Local_CDR, - #"BayesGDist_Local_FWR" = BayesGDist_Local_FWR)) -# "BayesGDist_Global_CDR" = BayesGDist_Global_CDR, -# "BayesGDist_Global_FWR" = BayesGDist_Global_FWR) ) - - -} - - -calculate_bayesG <- function( x=array(), N=array(), p=array(), max_sigma=20, length_sigma=4001){ - G <- max(length(x),length(N),length(p)) - x=array(x,dim=G) - N=array(N,dim=G) - p=array(p,dim=G) - - indexOfZero = N>0 & p>0 - N = N[indexOfZero] - x = x[indexOfZero] - p = p[indexOfZero] - G <- length(x) - - if(G){ - - cons<-array( dim=c(length_sigma,G) ) - if(G==1) { - return(calculate_bayes(x=x[G],N=N[G],p=p[G],max_sigma=max_sigma,length_sigma=length_sigma)) - } - else { - for(g in 1:G) cons[,g] <- calculate_bayes(x=x[g],N=N[g],p=p[g],max_sigma=max_sigma,length_sigma=length_sigma) - listMatG <- convolutionPowersOfTwoByTwos(cons,length_sigma=length_sigma) - y<-calculate_bayesGHelper(listMatG,length_sigma=length_sigma) - return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) - } - }else{ - return(NA) - } -} - - -calculate_bayesGHelper <- function( listMatG,length_sigma=4001 ){ - matG <- listMatG[[1]] - groups <- listMatG[[2]] - i = 1 - resConv <- matG[,i] - denom <- 2^groups[i] - if(length(groups)>1){ - while( i0)) ){ - -# ONEmerStartPos = 1:(seqLength) -# ONEmerLength <- length(ONEmerStartPos) - ONEmerGL <- s2c(seqGL) - ONEmerSeq <- s2c(seqInput) - - #Background - for(ONEmerIndex in 1:seqLength){ - ONEmer = ONEmerGL[ONEmerIndex] - if(ONEmer!="N"){ - ONEmerCodonPos = getCodonPos(ONEmerIndex) - ONEmerReadingFrameCodon = c2s(ONEmerGL[ONEmerCodonPos]) - ONEmerReadingFrameCodonInputSeq = c2s(ONEmerSeq[ONEmerCodonPos] ) - - # All mutations model - #if(!any(grep("N",ONEmerReadingFrameCodon))){ - if(model==0){ - if(stopMutations==0){ - if(!any(grep("N",ONEmerReadingFrameCodonInputSeq))) - BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + 1) - }else{ - if( !any(grep("N",ONEmerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)!="*"){ - positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex)#positionsWithinCodon[(ONEmerCodonPos[1]%%3)+1] - BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + probNonStopMutations[ONEmerReadingFrameCodon,positionWithinCodon]) - } - } - }else{ # Only silent mutations - if( !any(grep("N",ONEmerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)!="*" & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)==translateCodonToAminoAcid(ONEmerReadingFrameCodon) ){ - positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex) - BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + probSMutations[ONEmerReadingFrameCodon,positionWithinCodon]) - } - } - } - } - } - - #Mutations - if(stopMutations==1) mutationCount = mutationCount[mutationCount!="Stop"] - if(model==1) mutationCount = mutationCount[mutationCount=="S"] - mutationPositions = as.numeric(names(mutationCount)) - mutationCount = mutationCount[mutationPositions>2 & mutationPositions<(seqLength-1)] - mutationPositions = mutationPositions[mutationPositions>2 & mutationPositions<(seqLength-1)] - countMutations = 0 - for(mutationPosition in mutationPositions){ - ONEmerIndex = mutationPosition - ONEmer = ONEmerSeq[ONEmerIndex] - GLONEmer = ONEmerGL[ONEmerIndex] - ONEmerCodonPos = getCodonPos(ONEmerIndex) - ONEmerReadingFrameCodon = c2s(ONEmerSeq[ONEmerCodonPos]) - ONEmerReadingFrameCodonGL =c2s(ONEmerGL[ONEmerCodonPos]) - if(!any(grep("N",ONEmer)) & !any(grep("N",GLONEmer))){ - if(model==0){ - countMutations = countMutations + 1 - MutationMatrix[GLONEmer] <- (MutationMatrix[GLONEmer] + 1) - MutationCountMatrix[GLONEmer] <- (MutationCountMatrix[GLONEmer] + 1) - }else{ - if( translateCodonToAminoAcid(ONEmerReadingFrameCodonGL)!="*" ){ - countMutations = countMutations + 1 - positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex) - glNuc = substr(ONEmerReadingFrameCodonGL,positionWithinCodon,positionWithinCodon) - inputNuc = substr(ONEmerReadingFrameCodon,positionWithinCodon,positionWithinCodon) - MutationMatrix[GLONEmer] <- (MutationMatrix[GLONEmer] + substitution[glNuc,inputNuc]) - MutationCountMatrix[GLONEmer] <- (MutationCountMatrix[GLONEmer] + 1) - } - } - } - } - - seqMutability = MutationMatrix/BackgroundMatrix - seqMutability = seqMutability/sum(seqMutability,na.rm=TRUE) - #cat(inputMatrixIndex,"\t",countMutations,"\n") - return(list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix, "BackgroundMatrix"=BackgroundMatrix)) -# tmp<-list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix) - } - } - -################ -# $Id: trim.R 989 2006-10-29 15:28:26Z ggorjan $ - -trim <- function(s, recode.factor=TRUE, ...) - UseMethod("trim", s) - -trim.default <- function(s, recode.factor=TRUE, ...) - s - -trim.character <- function(s, recode.factor=TRUE, ...) -{ - s <- sub(pattern="^ +", replacement="", x=s) - s <- sub(pattern=" +$", replacement="", x=s) - s -} - -trim.factor <- function(s, recode.factor=TRUE, ...) -{ - levels(s) <- trim(levels(s)) - if(recode.factor) { - dots <- list(x=s, ...) - if(is.null(dots$sort)) dots$sort <- sort - s <- do.call(what=reorder.factor, args=dots) - } - s -} - -trim.list <- function(s, recode.factor=TRUE, ...) - lapply(s, trim, recode.factor=recode.factor, ...) - -trim.data.frame <- function(s, recode.factor=TRUE, ...) -{ - s[] <- trim.list(s, recode.factor=recode.factor, ...) - s -} -####################################### -# Compute the expected for each sequence-germline pair by codon -getExpectedIndividualByCodon <- function(matInput){ -if( any(grep("multicore",search())) ){ - facGL <- factor(matInput[,2]) - facLevels = levels(facGL) - LisGLs_MutabilityU = mclapply(1:length(facLevels), function(x){ - computeMutabilities(facLevels[x]) - }) - facIndex = match(facGL,facLevels) - - LisGLs_Mutability = mclapply(1:nrow(matInput), function(x){ - cInput = rep(NA,nchar(matInput[x,1])) - cInput[s2c(matInput[x,1])!="N"] = 1 - LisGLs_MutabilityU[[facIndex[x]]] * cInput - }) - - LisGLs_Targeting = mclapply(1:dim(matInput)[1], function(x){ - computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) - }) - - LisGLs_MutationTypes = mclapply(1:length(matInput[,2]),function(x){ - #print(x) - computeMutationTypes(matInput[x,2]) - }) - - LisGLs_R_Exp = mclapply(1:nrow(matInput), function(x){ - Exp_R <- rollapply(as.zoo(1:readEnd),width=3,by=3, - function(codonNucs){ - RPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="R") - sum( LisGLs_Targeting[[x]][,codonNucs][RPos], na.rm=T ) - } - ) - }) - - LisGLs_S_Exp = mclapply(1:nrow(matInput), function(x){ - Exp_S <- rollapply(as.zoo(1:readEnd),width=3,by=3, - function(codonNucs){ - SPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="S") - sum( LisGLs_Targeting[[x]][,codonNucs][SPos], na.rm=T ) - } - ) - }) - - Exp_R = matrix(unlist(LisGLs_R_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) - Exp_S = matrix(unlist(LisGLs_S_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) - return( list( "Expected_R"=Exp_R, "Expected_S"=Exp_S) ) - }else{ - facGL <- factor(matInput[,2]) - facLevels = levels(facGL) - LisGLs_MutabilityU = lapply(1:length(facLevels), function(x){ - computeMutabilities(facLevels[x]) - }) - facIndex = match(facGL,facLevels) - - LisGLs_Mutability = lapply(1:nrow(matInput), function(x){ - cInput = rep(NA,nchar(matInput[x,1])) - cInput[s2c(matInput[x,1])!="N"] = 1 - LisGLs_MutabilityU[[facIndex[x]]] * cInput - }) - - LisGLs_Targeting = lapply(1:dim(matInput)[1], function(x){ - computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) - }) - - LisGLs_MutationTypes = lapply(1:length(matInput[,2]),function(x){ - #print(x) - computeMutationTypes(matInput[x,2]) - }) - - LisGLs_R_Exp = lapply(1:nrow(matInput), function(x){ - Exp_R <- rollapply(as.zoo(1:readEnd),width=3,by=3, - function(codonNucs){ - RPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="R") - sum( LisGLs_Targeting[[x]][,codonNucs][RPos], na.rm=T ) - } - ) - }) - - LisGLs_S_Exp = lapply(1:nrow(matInput), function(x){ - Exp_S <- rollapply(as.zoo(1:readEnd),width=3,by=3, - function(codonNucs){ - SPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="S") - sum( LisGLs_Targeting[[x]][,codonNucs][SPos], na.rm=T ) - } - ) - }) - - Exp_R = matrix(unlist(LisGLs_R_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) - Exp_S = matrix(unlist(LisGLs_S_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) - return( list( "Expected_R"=Exp_R, "Expected_S"=Exp_S) ) - } -} - -# getObservedMutationsByCodon <- function(listMutations){ -# numbSeqs <- length(listMutations) -# obsMu_R <- matrix(0,nrow=numbSeqs,ncol=readEnd/3,dimnames=list(c(1:numbSeqs),c(1:(readEnd/3)))) -# obsMu_S <- obsMu_R -# temp <- mclapply(1:length(listMutations), function(i){ -# arrMutations = listMutations[[i]] -# RPos = as.numeric(names(arrMutations)[arrMutations=="R"]) -# RPos <- sapply(RPos,getCodonNumb) -# if(any(RPos)){ -# tabR <- table(RPos) -# obsMu_R[i,as.numeric(names(tabR))] <<- tabR -# } -# -# SPos = as.numeric(names(arrMutations)[arrMutations=="S"]) -# SPos <- sapply(SPos,getCodonNumb) -# if(any(SPos)){ -# tabS <- table(SPos) -# obsMu_S[i,names(tabS)] <<- tabS -# } -# } -# ) -# return( list( "Observed_R"=obsMu_R, "Observed_S"=obsMu_S) ) -# } - -getObservedMutationsByCodon <- function(listMutations){ - numbSeqs <- length(listMutations) - obsMu_R <- matrix(0,nrow=numbSeqs,ncol=readEnd/3,dimnames=list(c(1:numbSeqs),c(1:(readEnd/3)))) - obsMu_S <- obsMu_R - temp <- lapply(1:length(listMutations), function(i){ - arrMutations = listMutations[[i]] - RPos = as.numeric(names(arrMutations)[arrMutations=="R"]) - RPos <- sapply(RPos,getCodonNumb) - if(any(RPos)){ - tabR <- table(RPos) - obsMu_R[i,as.numeric(names(tabR))] <<- tabR - } - - SPos = as.numeric(names(arrMutations)[arrMutations=="S"]) - SPos <- sapply(SPos,getCodonNumb) - if(any(SPos)){ - tabS <- table(SPos) - obsMu_S[i,names(tabS)] <<- tabS - } - } - ) - return( list( "Observed_R"=obsMu_R, "Observed_S"=obsMu_S) ) -} - +######################################################################################### +# License Agreement +# +# THIS WORK IS PROVIDED UNDER THE TERMS OF THIS CREATIVE COMMONS PUBLIC LICENSE +# ("CCPL" OR "LICENSE"). THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER +# APPLICABLE LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UNDER THIS LICENSE +# OR COPYRIGHT LAW IS PROHIBITED. +# +# BY EXERCISING ANY RIGHTS TO THE WORK PROVIDED HERE, YOU ACCEPT AND AGREE TO BE +# BOUND BY THE TERMS OF THIS LICENSE. TO THE EXTENT THIS LICENSE MAY BE CONSIDERED +# TO BE A CONTRACT, THE LICENSOR GRANTS YOU THE RIGHTS CONTAINED HERE IN +# CONSIDERATION OF YOUR ACCEPTANCE OF SUCH TERMS AND CONDITIONS. +# +# BASELIne: Bayesian Estimation of Antigen-Driven Selection in Immunoglobulin Sequences +# Coded by: Mohamed Uduman & Gur Yaari +# Copyright 2012 Kleinstein Lab +# Version: 1.3 (01/23/2014) +######################################################################################### + +# Global variables + + FILTER_BY_MUTATIONS = 1000 + + # Nucleotides + NUCLEOTIDES = c("A","C","G","T") + + # Amino Acids + AMINO_ACIDS <- c("F", "F", "L", "L", "S", "S", "S", "S", "Y", "Y", "*", "*", "C", "C", "*", "W", "L", "L", "L", "L", "P", "P", "P", "P", "H", "H", "Q", "Q", "R", "R", "R", "R", "I", "I", "I", "M", "T", "T", "T", "T", "N", "N", "K", "K", "S", "S", "R", "R", "V", "V", "V", "V", "A", "A", "A", "A", "D", "D", "E", "E", "G", "G", "G", "G") + names(AMINO_ACIDS) <- c("TTT", "TTC", "TTA", "TTG", "TCT", "TCC", "TCA", "TCG", "TAT", "TAC", "TAA", "TAG", "TGT", "TGC", "TGA", "TGG", "CTT", "CTC", "CTA", "CTG", "CCT", "CCC", "CCA", "CCG", "CAT", "CAC", "CAA", "CAG", "CGT", "CGC", "CGA", "CGG", "ATT", "ATC", "ATA", "ATG", "ACT", "ACC", "ACA", "ACG", "AAT", "AAC", "AAA", "AAG", "AGT", "AGC", "AGA", "AGG", "GTT", "GTC", "GTA", "GTG", "GCT", "GCC", "GCA", "GCG", "GAT", "GAC", "GAA", "GAG", "GGT", "GGC", "GGA", "GGG") + names(AMINO_ACIDS) <- names(AMINO_ACIDS) + + #Amino Acid Traits + #"*" "A" "C" "D" "E" "F" "G" "H" "I" "K" "L" "M" "N" "P" "Q" "R" "S" "T" "V" "W" "Y" + #B = "Hydrophobic/Burried" N = "Intermediate/Neutral" S="Hydrophilic/Surface") + TRAITS_AMINO_ACIDS_CHOTHIA98 <- c("*","N","B","S","S","B","N","N","B","S","B","B","S","N","S","S","N","N","B","B","N") + names(TRAITS_AMINO_ACIDS_CHOTHIA98) <- sort(unique(AMINO_ACIDS)) + TRAITS_AMINO_ACIDS <- array(NA,21) + + # Codon Table + CODON_TABLE <- as.data.frame(matrix(NA,ncol=64,nrow=12)) + + # Substitution Model: Smith DS et al. 1996 + substitution_Literature_Mouse <- matrix(c(0, 0.156222928, 0.601501588, 0.242275484, 0.172506739, 0, 0.241239892, 0.586253369, 0.54636291, 0.255795364, 0, 0.197841727, 0.290240811, 0.467680608, 0.24207858, 0),nrow=4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) + substitution_Flu_Human <- matrix(c(0,0.2795596,0.5026927,0.2177477,0.1693210,0,0.3264723,0.5042067,0.4983549,0.3328321,0,0.1688130,0.2021079,0.4696077,0.3282844,0),4,4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) + substitution_Flu25_Human <- matrix(c(0,0.2580641,0.5163685,0.2255674,0.1541125,0,0.3210224,0.5248651,0.5239281,0.3101292,0,0.1659427,0.1997207,0.4579444,0.3423350,0),4,4,byrow=T,dimnames=list(NUCLEOTIDES,NUCLEOTIDES)) + load("FiveS_Substitution.RData") + + # Mutability Models: Shapiro GS et al. 2002 + triMutability_Literature_Human <- matrix(c(0.24, 1.2, 0.96, 0.43, 2.14, 2, 1.11, 1.9, 0.85, 1.83, 2.36, 1.31, 0.82, 0.52, 0.89, 1.33, 1.4, 0.82, 1.83, 0.73, 1.83, 1.62, 1.53, 0.57, 0.92, 0.42, 0.42, 1.47, 3.44, 2.58, 1.18, 0.47, 0.39, 1.12, 1.8, 0.68, 0.47, 2.19, 2.35, 2.19, 1.05, 1.84, 1.26, 0.28, 0.98, 2.37, 0.66, 1.58, 0.67, 0.92, 1.76, 0.83, 0.97, 0.56, 0.75, 0.62, 2.26, 0.62, 0.74, 1.11, 1.16, 0.61, 0.88, 0.67, 0.37, 0.07, 1.08, 0.46, 0.31, 0.94, 0.62, 0.57, 0.29, NA, 1.44, 0.46, 0.69, 0.57, 0.24, 0.37, 1.1, 0.99, 1.39, 0.6, 2.26, 1.24, 1.36, 0.52, 0.33, 0.26, 1.25, 0.37, 0.58, 1.03, 1.2, 0.34, 0.49, 0.33, 2.62, 0.16, 0.4, 0.16, 0.35, 0.75, 1.85, 0.94, 1.61, 0.85, 2.09, 1.39, 0.3, 0.52, 1.33, 0.29, 0.51, 0.26, 0.51, 3.83, 2.01, 0.71, 0.58, 0.62, 1.07, 0.28, 1.2, 0.74, 0.25, 0.59, 1.09, 0.91, 1.36, 0.45, 2.89, 1.27, 3.7, 0.69, 0.28, 0.41, 1.17, 0.56, 0.93, 3.41, 1, 1, NA, 5.9, 0.74, 2.51, 2.24, 2.24, 1.95, 3.32, 2.34, 1.3, 2.3, 1, 0.66, 0.73, 0.93, 0.41, 0.65, 0.89, 0.65, 0.32, NA, 0.43, 0.85, 0.43, 0.31, 0.31, 0.23, 0.29, 0.57, 0.71, 0.48, 0.44, 0.76, 0.51, 1.7, 0.85, 0.74, 2.23, 2.08, 1.16, 0.51, 0.51, 1, 0.5, NA, NA, 0.71, 2.14), nrow=64,byrow=T) + triMutability_Literature_Mouse <- matrix(c(1.31, 1.35, 1.42, 1.18, 2.02, 2.02, 1.02, 1.61, 1.99, 1.42, 2.01, 1.03, 2.02, 0.97, 0.53, 0.71, 1.19, 0.83, 0.96, 0.96, 0, 1.7, 2.22, 0.59, 1.24, 1.07, 0.51, 1.68, 3.36, 3.36, 1.14, 0.29, 0.33, 0.9, 1.11, 0.63, 1.08, 2.07, 2.27, 1.74, 0.22, 1.19, 2.37, 1.15, 1.15, 1.56, 0.81, 0.34, 0.87, 0.79, 2.13, 0.49, 0.85, 0.97, 0.36, 0.82, 0.66, 0.63, 1.15, 0.94, 0.85, 0.25, 0.93, 1.19, 0.4, 0.2, 0.44, 0.44, 0.88, 1.06, 0.77, 0.39, 0, 0, 0, 0, 0, 0, 0.43, 0.43, 0.86, 0.59, 0.59, 0, 1.18, 0.86, 2.9, 1.66, 0.4, 0.2, 1.54, 0.43, 0.69, 1.71, 0.68, 0.55, 0.91, 0.7, 1.71, 0.09, 0.27, 0.63, 0.2, 0.45, 1.01, 1.63, 0.96, 1.48, 2.18, 1.2, 1.31, 0.66, 2.13, 0.49, 0, 0, 0, 2.97, 2.8, 0.79, 0.4, 0.5, 0.4, 0.11, 1.68, 0.42, 0.13, 0.44, 0.93, 0.71, 1.11, 1.19, 2.71, 1.08, 3.43, 0.4, 0.67, 0.47, 1.02, 0.14, 1.56, 1.98, 0.53, 0.33, 0.63, 2.06, 1.77, 1.46, 3.74, 2.93, 2.1, 2.18, 0.78, 0.73, 2.93, 0.63, 0.57, 0.17, 0.85, 0.52, 0.31, 0.31, 0, 0, 0.51, 0.29, 0.83, 0.54, 0.28, 0.47, 0.9, 0.99, 1.24, 2.47, 0.73, 0.23, 1.13, 0.24, 2.12, 0.24, 0.33, 0.83, 1.41, 0.62, 0.28, 0.35, 0.77, 0.17, 0.72, 0.58, 0.45, 0.41), nrow=64,byrow=T) + triMutability_Names <- c("AAA", "AAC", "AAG", "AAT", "ACA", "ACC", "ACG", "ACT", "AGA", "AGC", "AGG", "AGT", "ATA", "ATC", "ATG", "ATT", "CAA", "CAC", "CAG", "CAT", "CCA", "CCC", "CCG", "CCT", "CGA", "CGC", "CGG", "CGT", "CTA", "CTC", "CTG", "CTT", "GAA", "GAC", "GAG", "GAT", "GCA", "GCC", "GCG", "GCT", "GGA", "GGC", "GGG", "GGT", "GTA", "GTC", "GTG", "GTT", "TAA", "TAC", "TAG", "TAT", "TCA", "TCC", "TCG", "TCT", "TGA", "TGC", "TGG", "TGT", "TTA", "TTC", "TTG", "TTT") + load("FiveS_Mutability.RData") + +# Functions + + # Translate codon to amino acid + translateCodonToAminoAcid<-function(Codon){ + return(AMINO_ACIDS[Codon]) + } + + # Translate amino acid to trait change + translateAminoAcidToTraitChange<-function(AminoAcid){ + return(TRAITS_AMINO_ACIDS[AminoAcid]) + } + + # Initialize Amino Acid Trait Changes + initializeTraitChange <- function(traitChangeModel=1,species=1,traitChangeFileName=NULL){ + if(!is.null(traitChangeFileName)){ + tryCatch( + traitChange <- read.delim(traitChangeFileName,sep="\t",header=T) + , error = function(ex){ + cat("Error|Error reading trait changes. Please check file name/path and format.\n") + q() + } + ) + }else{ + traitChange <- TRAITS_AMINO_ACIDS_CHOTHIA98 + } + TRAITS_AMINO_ACIDS <<- traitChange + } + + # Read in formatted nucleotide substitution matrix + initializeSubstitutionMatrix <- function(substitutionModel,species,subsMatFileName=NULL){ + if(!is.null(subsMatFileName)){ + tryCatch( + subsMat <- read.delim(subsMatFileName,sep="\t",header=T) + , error = function(ex){ + cat("Error|Error reading substitution matrix. Please check file name/path and format.\n") + q() + } + ) + if(sum(apply(subsMat,1,sum)==1)!=4) subsMat = t(apply(subsMat,1,function(x)x/sum(x))) + }else{ + if(substitutionModel==1)subsMat <- substitution_Literature_Mouse + if(substitutionModel==2)subsMat <- substitution_Flu_Human + if(substitutionModel==3)subsMat <- substitution_Flu25_Human + + } + + if(substitutionModel==0){ + subsMat <- matrix(1,4,4) + subsMat[,] = 1/3 + subsMat[1,1] = 0 + subsMat[2,2] = 0 + subsMat[3,3] = 0 + subsMat[4,4] = 0 + } + + + NUCLEOTIDESN = c(NUCLEOTIDES,"N", "-") + if(substitutionModel==5){ + subsMat <- FiveS_Substitution + return(subsMat) + }else{ + subsMat <- rbind(subsMat,rep(NA,4),rep(NA,4)) + return( matrix(data.matrix(subsMat),6,4,dimnames=list(NUCLEOTIDESN,NUCLEOTIDES) ) ) + } + } + + + # Read in formatted Mutability file + initializeMutabilityMatrix <- function(mutabilityModel=1, species=1,mutabilityMatFileName=NULL){ + if(!is.null(mutabilityMatFileName)){ + tryCatch( + mutabilityMat <- read.delim(mutabilityMatFileName,sep="\t",header=T) + , error = function(ex){ + cat("Error|Error reading mutability matrix. Please check file name/path and format.\n") + q() + } + ) + }else{ + mutabilityMat <- triMutability_Literature_Human + if(species==2) mutabilityMat <- triMutability_Literature_Mouse + } + + if(mutabilityModel==0){ mutabilityMat <- matrix(1,64,3)} + + if(mutabilityModel==5){ + mutabilityMat <- FiveS_Mutability + return(mutabilityMat) + }else{ + return( matrix( data.matrix(mutabilityMat), 64, 3, dimnames=list(triMutability_Names,1:3)) ) + } + } + + # Read FASTA file formats + # Modified from read.fasta from the seqinR package + baseline.read.fasta <- + function (file = system.file("sequences/sample.fasta", package = "seqinr"), + seqtype = c("DNA", "AA"), as.string = FALSE, forceDNAtolower = TRUE, + set.attributes = TRUE, legacy.mode = TRUE, seqonly = FALSE, + strip.desc = FALSE, sizeof.longlong = .Machine$sizeof.longlong, + endian = .Platform$endian, apply.mask = TRUE) + { + seqtype <- match.arg(seqtype) + + lines <- readLines(file) + + if (legacy.mode) { + comments <- grep("^;", lines) + if (length(comments) > 0) + lines <- lines[-comments] + } + + + ind_groups<-which(substr(lines, 1L, 3L) == ">>>") + lines_mod<-lines + + if(!length(ind_groups)){ + lines_mod<-c(">>>All sequences combined",lines) + } + + ind_groups<-which(substr(lines_mod, 1L, 3L) == ">>>") + + lines <- array("BLA",dim=(length(ind_groups)+length(lines_mod))) + id<-sapply(1:length(ind_groups),function(i)ind_groups[i]+i-1)+1 + lines[id] <- "THIS IS A FAKE SEQUENCE" + lines[-id] <- lines_mod + rm(lines_mod) + + ind <- which(substr(lines, 1L, 1L) == ">") + nseq <- length(ind) + if (nseq == 0) { + stop("no line starting with a > character found") + } + start <- ind + 1 + end <- ind - 1 + + while( any(which(ind%in%end)) ){ + ind=ind[-which(ind%in%end)] + nseq <- length(ind) + if (nseq == 0) { + stop("no line starting with a > character found") + } + start <- ind + 1 + end <- ind - 1 + } + + end <- c(end[-1], length(lines)) + sequences <- lapply(seq_len(nseq), function(i) paste(lines[start[i]:end[i]], collapse = "")) + if (seqonly) + return(sequences) + nomseq <- lapply(seq_len(nseq), function(i) { + + #firstword <- strsplit(lines[ind[i]], " ")[[1]][1] + substr(lines[ind[i]], 2, nchar(lines[ind[i]])) + + }) + if (seqtype == "DNA") { + if (forceDNAtolower) { + sequences <- as.list(tolower(chartr(".","-",sequences))) + }else{ + sequences <- as.list(toupper(chartr(".","-",sequences))) + } + } + if (as.string == FALSE) + sequences <- lapply(sequences, s2c) + if (set.attributes) { + for (i in seq_len(nseq)) { + Annot <- lines[ind[i]] + if (strip.desc) + Annot <- substr(Annot, 2L, nchar(Annot)) + attributes(sequences[[i]]) <- list(name = nomseq[[i]], + Annot = Annot, class = switch(seqtype, AA = "SeqFastaAA", + DNA = "SeqFastadna")) + } + } + names(sequences) <- nomseq + return(sequences) + } + + + # Replaces non FASTA characters in input files with N + replaceNonFASTAChars <-function(inSeq="ACGTN-AApA"){ + gsub('[^ACGTNacgt[:punct:]-[:punct:].]','N',inSeq,perl=TRUE) + } + + # Find the germlines in the FASTA list + germlinesInFile <- function(seqIDs){ + firstChar = sapply(seqIDs,function(x){substr(x,1,1)}) + secondChar = sapply(seqIDs,function(x){substr(x,2,2)}) + return(firstChar==">" & secondChar!=">") + } + + # Find the groups in the FASTA list + groupsInFile <- function(seqIDs){ + sapply(seqIDs,function(x){substr(x,1,2)})==">>" + } + + # In the process of finding germlines/groups, expand from the start to end of the group + expandTillNext <- function(vecPosToID){ + IDs = names(vecPosToID) + posOfInterests = which(vecPosToID) + + expandedID = rep(NA,length(IDs)) + expandedIDNames = gsub(">","",IDs[posOfInterests]) + startIndexes = c(1,posOfInterests[-1]) + stopIndexes = c(posOfInterests[-1]-1,length(IDs)) + expandedID = unlist(sapply(1:length(startIndexes),function(i){ + rep(i,stopIndexes[i]-startIndexes[i]+1) + })) + names(expandedID) = unlist(sapply(1:length(startIndexes),function(i){ + rep(expandedIDNames[i],stopIndexes[i]-startIndexes[i]+1) + })) + return(expandedID) + } + + # Process FASTA (list) to return a matrix[input, germline) + processInputAdvanced <- function(inputFASTA){ + + seqIDs = names(inputFASTA) + numbSeqs = length(seqIDs) + posGermlines1 = germlinesInFile(seqIDs) + numbGermlines = sum(posGermlines1) + posGroups1 = groupsInFile(seqIDs) + numbGroups = sum(posGroups1) + consDef = NA + + if(numbGermlines==0){ + posGermlines = 2 + numbGermlines = 1 + } + + glPositionsSum = cumsum(posGermlines1) + glPositions = table(glPositionsSum) + #Find the position of the conservation row + consDefPos = as.numeric(names(glPositions[names(glPositions)!=0 & glPositions==1]))+1 + if( length(consDefPos)> 0 ){ + consDefID = match(consDefPos, glPositionsSum) + #The coservation rows need to be pulled out and stores seperately + consDef = inputFASTA[consDefID] + inputFASTA = inputFASTA[-consDefID] + + seqIDs = names(inputFASTA) + numbSeqs = length(seqIDs) + posGermlines1 = germlinesInFile(seqIDs) + numbGermlines = sum(posGermlines1) + posGroups1 = groupsInFile(seqIDs) + numbGroups = sum(posGroups1) + if(numbGermlines==0){ + posGermlines = 2 + numbGermlines = 1 + } + } + + posGroups <- expandTillNext(posGroups1) + posGermlines <- expandTillNext(posGermlines1) + posGermlines[posGroups1] = 0 + names(posGermlines)[posGroups1] = names(posGroups)[posGroups1] + posInput = rep(TRUE,numbSeqs) + posInput[posGroups1 | posGermlines1] = FALSE + + matInput = matrix(NA, nrow=sum(posInput), ncol=2) + rownames(matInput) = seqIDs[posInput] + colnames(matInput) = c("Input","Germline") + + vecInputFASTA = unlist(inputFASTA) + matInput[,1] = vecInputFASTA[posInput] + matInput[,2] = vecInputFASTA[ which( names(inputFASTA)%in%paste(">",names(posGermlines)[posInput],sep="") )[ posGermlines[posInput]] ] + + germlines = posGermlines[posInput] + groups = posGroups[posInput] + + return( list("matInput"=matInput, "germlines"=germlines, "groups"=groups, "conservationDefinition"=consDef )) + } + + + # Replace leading and trailing dashes in the sequence + replaceLeadingTrailingDashes <- function(x,readEnd){ + iiGap = unlist(gregexpr("-",x[1])) + ggGap = unlist(gregexpr("-",x[2])) + #posToChange = intersect(iiGap,ggGap) + + + seqIn = replaceLeadingTrailingDashesHelper(x[1]) + seqGL = replaceLeadingTrailingDashesHelper(x[2]) + seqTemplate = rep('N',readEnd) + seqIn <- c(seqIn,seqTemplate[(length(seqIn)+1):readEnd]) + seqGL <- c(seqGL,seqTemplate[(length(seqGL)+1):readEnd]) +# if(posToChange!=-1){ +# seqIn[posToChange] = "-" +# seqGL[posToChange] = "-" +# } + + seqIn = c2s(seqIn[1:readEnd]) + seqGL = c2s(seqGL[1:readEnd]) + + lenGL = nchar(seqGL) + if(lenGL seqLen ) + trimmedSeq = substr(seqToTrim,1, ( (getCodonPos(seqLen)[1])-1 ) ) + + return(trimmedSeq) + } + + # Given a nuclotide position, returns the pos of the 3 nucs that made the codon + # e.g. nuc 86 is part of nucs 85,86,87 + getCodonPos <- function(nucPos){ + codonNum = (ceiling(nucPos/3))*3 + return( (codonNum-2):codonNum) + } + + # Given a nuclotide position, returns the codon number + # e.g. nuc 86 = codon 29 + getCodonNumb <- function(nucPos){ + return( ceiling(nucPos/3) ) + } + + # Given a codon, returns all the nuc positions that make the codon + getCodonNucs <- function(codonNumb){ + getCodonPos(codonNumb*3) + } + + computeCodonTable <- function(testID=1){ + + if(testID<=4){ + # Pre-compute every codons + intCounter = 1 + for(pOne in NUCLEOTIDES){ + for(pTwo in NUCLEOTIDES){ + for(pThree in NUCLEOTIDES){ + codon = paste(pOne,pTwo,pThree,sep="") + colnames(CODON_TABLE)[intCounter] = codon + intCounter = intCounter + 1 + CODON_TABLE[,codon] = mutationTypeOptimized(cbind(permutateAllCodon(codon),rep(codon,12))) + } + } + } + chars = c("N","A","C","G","T", "-") + for(a in chars){ + for(b in chars){ + for(c in chars){ + if(a=="N" | b=="N" | c=="N"){ + #cat(paste(a,b,c),sep="","\n") + CODON_TABLE[,paste(a,b,c,sep="")] = rep(NA,12) + } + } + } + } + + chars = c("-","A","C","G","T") + for(a in chars){ + for(b in chars){ + for(c in chars){ + if(a=="-" | b=="-" | c=="-"){ + #cat(paste(a,b,c),sep="","\n") + CODON_TABLE[,paste(a,b,c,sep="")] = rep(NA,12) + } + } + } + } + CODON_TABLE <<- as.matrix(CODON_TABLE) + } + } + + collapseClone <- function(vecInputSeqs,glSeq,readEnd,nonTerminalOnly=0){ + #print(length(vecInputSeqs)) + vecInputSeqs = unique(vecInputSeqs) + if(length(vecInputSeqs)==1){ + return( list( c(vecInputSeqs,glSeq), F) ) + }else{ + charInputSeqs <- sapply(vecInputSeqs, function(x){ + s2c(x)[1:readEnd] + }) + charGLSeq <- s2c(glSeq) + matClone <- sapply(1:readEnd, function(i){ + posNucs = unique(charInputSeqs[i,]) + posGL = charGLSeq[i] + error = FALSE + if(posGL=="-" & sum(!(posNucs%in%c("-","N")))==0 ){ + return(c("-",error)) + } + if(length(posNucs)==1) + return(c(posNucs[1],error)) + else{ + if("N"%in%posNucs){ + error=TRUE + } + if(sum(!posNucs[posNucs!="N"]%in%posGL)==0){ + return( c(posGL,error) ) + }else{ + #return( c(sample(posNucs[posNucs!="N"],1),error) ) + if(nonTerminalOnly==0){ + return( c(sample(charInputSeqs[i,charInputSeqs[i,]!="N" & charInputSeqs[i,]!=posGL],1),error) ) + }else{ + posNucs = charInputSeqs[i,charInputSeqs[i,]!="N" & charInputSeqs[i,]!=posGL] + posNucsTable = table(posNucs) + if(sum(posNucsTable>1)==0){ + return( c(posGL,error) ) + }else{ + return( c(sample( posNucs[posNucs%in%names(posNucsTable)[posNucsTable>1]],1),error) ) + } + } + + } + } + }) + + + #print(length(vecInputSeqs)) + return(list(c(c2s(matClone[1,]),glSeq),"TRUE"%in%matClone[2,])) + } + } + + # Compute the expected for each sequence-germline pair + getExpectedIndividual <- function(matInput){ + if( any(grep("multicore",search())) ){ + facGL <- factor(matInput[,2]) + facLevels = levels(facGL) + LisGLs_MutabilityU = mclapply(1:length(facLevels), function(x){ + computeMutabilities(facLevels[x]) + }) + facIndex = match(facGL,facLevels) + + LisGLs_Mutability = mclapply(1:nrow(matInput), function(x){ + cInput = rep(NA,nchar(matInput[x,1])) + cInput[s2c(matInput[x,1])!="N"] = 1 + LisGLs_MutabilityU[[facIndex[x]]] * cInput + }) + + LisGLs_Targeting = mclapply(1:dim(matInput)[1], function(x){ + computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) + }) + + LisGLs_MutationTypes = mclapply(1:length(matInput[,2]),function(x){ + #print(x) + computeMutationTypes(matInput[x,2]) + }) + + LisGLs_Exp = mclapply(1:dim(matInput)[1], function(x){ + computeExpected(LisGLs_Targeting[[x]],LisGLs_MutationTypes[[x]]) + }) + + ul_LisGLs_Exp = unlist(LisGLs_Exp) + return(matrix(ul_LisGLs_Exp,ncol=4,nrow=(length(ul_LisGLs_Exp)/4),byrow=T)) + }else{ + facGL <- factor(matInput[,2]) + facLevels = levels(facGL) + LisGLs_MutabilityU = lapply(1:length(facLevels), function(x){ + computeMutabilities(facLevels[x]) + }) + facIndex = match(facGL,facLevels) + + LisGLs_Mutability = lapply(1:nrow(matInput), function(x){ + cInput = rep(NA,nchar(matInput[x,1])) + cInput[s2c(matInput[x,1])!="N"] = 1 + LisGLs_MutabilityU[[facIndex[x]]] * cInput + }) + + LisGLs_Targeting = lapply(1:dim(matInput)[1], function(x){ + computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) + }) + + LisGLs_MutationTypes = lapply(1:length(matInput[,2]),function(x){ + #print(x) + computeMutationTypes(matInput[x,2]) + }) + + LisGLs_Exp = lapply(1:dim(matInput)[1], function(x){ + computeExpected(LisGLs_Targeting[[x]],LisGLs_MutationTypes[[x]]) + }) + + ul_LisGLs_Exp = unlist(LisGLs_Exp) + return(matrix(ul_LisGLs_Exp,ncol=4,nrow=(length(ul_LisGLs_Exp)/4),byrow=T)) + + } + } + + # Compute mutabilities of sequence based on the tri-nucleotide model + computeMutabilities <- function(paramSeq){ + seqLen = nchar(paramSeq) + seqMutabilites = rep(NA,seqLen) + + gaplessSeq = gsub("-", "", paramSeq) + gaplessSeqLen = nchar(gaplessSeq) + gaplessSeqMutabilites = rep(NA,gaplessSeqLen) + + if(mutabilityModel!=5){ + pos<- 3:(gaplessSeqLen) + subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) + gaplessSeqMutabilites[pos] = + tapply( c( + getMutability( substr(subSeq,1,3), 3) , + getMutability( substr(subSeq,2,4), 2), + getMutability( substr(subSeq,3,5), 1) + ),rep(1:(gaplessSeqLen-2),3),mean,na.rm=TRUE + ) + #Pos 1 + subSeq = substr(gaplessSeq,1,3) + gaplessSeqMutabilites[1] = getMutability(subSeq , 1) + #Pos 2 + subSeq = substr(gaplessSeq,1,4) + gaplessSeqMutabilites[2] = mean( c( + getMutability( substr(subSeq,1,3), 2) , + getMutability( substr(subSeq,2,4), 1) + ),na.rm=T + ) + seqMutabilites[which(s2c(paramSeq)!="-")]<- gaplessSeqMutabilites + return(seqMutabilites) + }else{ + + pos<- 3:(gaplessSeqLen) + subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) + gaplessSeqMutabilites[pos] = sapply(subSeq,function(x){ getMutability5(x) }, simplify=T) + seqMutabilites[which(s2c(paramSeq)!="-")]<- gaplessSeqMutabilites + return(seqMutabilites) + } + + } + + # Returns the mutability of a triplet at a given position + getMutability <- function(codon, pos=1:3){ + triplets <- rownames(mutability) + mutability[ match(codon,triplets) ,pos] + } + + getMutability5 <- function(fivemer){ + return(mutability[fivemer]) + } + + # Returns the substitution probabilty + getTransistionProb <- function(nuc){ + substitution[nuc,] + } + + getTransistionProb5 <- function(fivemer){ + if(any(which(fivemer==colnames(substitution)))){ + return(substitution[,fivemer]) + }else{ + return(array(NA,4)) + } + } + + # Given a nuc, returns the other 3 nucs it can mutate to + canMutateTo <- function(nuc){ + NUCLEOTIDES[- which(NUCLEOTIDES==nuc)] + } + + # Given a nucleotide, returns the probabilty of other nucleotide it can mutate to + canMutateToProb <- function(nuc){ + substitution[nuc,canMutateTo(nuc)] + } + + # Compute targeting, based on precomputed mutatbility & substitution + computeTargeting <- function(param_strSeq,param_vecMutabilities){ + + if(substitutionModel!=5){ + vecSeq = s2c(param_strSeq) + matTargeting = sapply( 1:length(vecSeq), function(x) { param_vecMutabilities[x] * getTransistionProb(vecSeq[x]) } ) + #matTargeting = apply( rbind(vecSeq,param_vecMutabilities),2, function(x) { as.vector(as.numeric(x[2]) * getTransistionProb(x[1])) } ) + dimnames( matTargeting ) = list(NUCLEOTIDES,1:(length(vecSeq))) + return (matTargeting) + }else{ + + seqLen = nchar(param_strSeq) + seqsubstitution = matrix(NA,ncol=seqLen,nrow=4) + paramSeq <- param_strSeq + gaplessSeq = gsub("-", "", paramSeq) + gaplessSeqLen = nchar(gaplessSeq) + gaplessSeqSubstitution = matrix(NA,ncol=gaplessSeqLen,nrow=4) + + pos<- 3:(gaplessSeqLen) + subSeq = substr(rep(gaplessSeq,gaplessSeqLen-2),(pos-2),(pos+2)) + gaplessSeqSubstitution[,pos] = sapply(subSeq,function(x){ getTransistionProb5(x) }, simplify=T) + seqsubstitution[,which(s2c(paramSeq)!="-")]<- gaplessSeqSubstitution + #matTargeting <- param_vecMutabilities %*% seqsubstitution + matTargeting <- sweep(seqsubstitution,2,param_vecMutabilities,`*`) + dimnames( matTargeting ) = list(NUCLEOTIDES,1:(seqLen)) + return (matTargeting) + } + } + + # Compute the mutations types + computeMutationTypes <- function(param_strSeq){ + #cat(param_strSeq,"\n") + #vecSeq = trimToLastCodon(param_strSeq) + lenSeq = nchar(param_strSeq) + vecCodons = sapply({1:(lenSeq/3)}*3-2,function(x){substr(param_strSeq,x,x+2)}) + matMutationTypes = matrix( unlist(CODON_TABLE[,vecCodons]) ,ncol=lenSeq,nrow=4, byrow=F) + dimnames( matMutationTypes ) = list(NUCLEOTIDES,1:(ncol(matMutationTypes))) + return(matMutationTypes) + } + computeMutationTypesFast <- function(param_strSeq){ + matMutationTypes = matrix( CODON_TABLE[,param_strSeq] ,ncol=3,nrow=4, byrow=F) + #dimnames( matMutationTypes ) = list(NUCLEOTIDES,1:(length(vecSeq))) + return(matMutationTypes) + } + mutationTypeOptimized <- function( matOfCodons ){ + apply( matOfCodons,1,function(x){ mutationType(x[2],x[1]) } ) + } + + # Returns a vector of codons 1 mutation away from the given codon + permutateAllCodon <- function(codon){ + cCodon = s2c(codon) + matCodons = t(array(cCodon,dim=c(3,12))) + matCodons[1:4,1] = NUCLEOTIDES + matCodons[5:8,2] = NUCLEOTIDES + matCodons[9:12,3] = NUCLEOTIDES + apply(matCodons,1,c2s) + } + + # Given two codons, tells you if the mutation is R or S (based on your definition) + mutationType <- function(codonFrom,codonTo){ + if(testID==4){ + if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ + return(NA) + }else{ + mutationType = "S" + if( translateAminoAcidToTraitChange(translateCodonToAminoAcid(codonFrom)) != translateAminoAcidToTraitChange(translateCodonToAminoAcid(codonTo)) ){ + mutationType = "R" + } + if(translateCodonToAminoAcid(codonTo)=="*" | translateCodonToAminoAcid(codonFrom)=="*"){ + mutationType = "Stop" + } + return(mutationType) + } + }else if(testID==5){ + if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ + return(NA) + }else{ + if(codonFrom==codonTo){ + mutationType = "S" + }else{ + codonFrom = s2c(codonFrom) + codonTo = s2c(codonTo) + mutationType = "Stop" + nucOfI = codonFrom[which(codonTo!=codonFrom)] + if(nucOfI=="C"){ + mutationType = "R" + }else if(nucOfI=="G"){ + mutationType = "S" + } + } + return(mutationType) + } + }else{ + if( is.na(codonFrom) | is.na(codonTo) | is.na(translateCodonToAminoAcid(codonFrom)) | is.na(translateCodonToAminoAcid(codonTo)) ){ + return(NA) + }else{ + mutationType = "S" + if( translateCodonToAminoAcid(codonFrom) != translateCodonToAminoAcid(codonTo) ){ + mutationType = "R" + } + if(translateCodonToAminoAcid(codonTo)=="*" | translateCodonToAminoAcid(codonFrom)=="*"){ + mutationType = "Stop" + } + return(mutationType) + } + } + } + + + #given a mat of targeting & it's corresponding mutationtypes returns + #a vector of Exp_RCDR,Exp_SCDR,Exp_RFWR,Exp_RFWR + computeExpected <- function(paramTargeting,paramMutationTypes){ + # Replacements + RPos = which(paramMutationTypes=="R") + #FWR + Exp_R_FWR = sum(paramTargeting[ RPos[which(FWR_Nuc_Mat[RPos]==T)] ],na.rm=T) + #CDR + Exp_R_CDR = sum(paramTargeting[ RPos[which(CDR_Nuc_Mat[RPos]==T)] ],na.rm=T) + # Silents + SPos = which(paramMutationTypes=="S") + #FWR + Exp_S_FWR = sum(paramTargeting[ SPos[which(FWR_Nuc_Mat[SPos]==T)] ],na.rm=T) + #CDR + Exp_S_CDR = sum(paramTargeting[ SPos[which(CDR_Nuc_Mat[SPos]==T)] ],na.rm=T) + + return(c(Exp_R_CDR,Exp_S_CDR,Exp_R_FWR,Exp_S_FWR)) + } + + # Count the mutations in a sequence + # each mutation is treated independently + analyzeMutations2NucUri_website <- function( rev_in_matrix ){ + paramGL = rev_in_matrix[2,] + paramSeq = rev_in_matrix[1,] + + #Fill seq with GL seq if gapped + #if( any(paramSeq=="-") ){ + # gapPos_Seq = which(paramSeq=="-") + # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "-"] + # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] + #} + + + #if( any(paramSeq=="N") ){ + # gapPos_Seq = which(paramSeq=="N") + # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] + # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] + #} + + analyzeMutations2NucUri( matrix(c( paramGL, paramSeq ),2,length(paramGL),byrow=T) ) + + } + + #1 = GL + #2 = Seq + analyzeMutations2NucUri <- function( in_matrix=matrix(c(c("A","A","A","C","C","C"),c("A","G","G","C","C","A")),2,6,byrow=T) ){ + paramGL = in_matrix[2,] + paramSeq = in_matrix[1,] + paramSeqUri = paramGL + #mutations = apply(rbind(paramGL,paramSeq), 2, function(x){!x[1]==x[2]}) + mutations_val = paramGL != paramSeq + if(any(mutations_val)){ + mutationPos = {1:length(mutations_val)}[mutations_val] + mutationPos = mutationPos[sapply(mutationPos, function(x){!any(paramSeq[getCodonPos(x)]=="N")})] + length_mutations =length(mutationPos) + mutationInfo = rep(NA,length_mutations) + if(any(mutationPos)){ + + pos<- mutationPos + pos_array<-array(sapply(pos,getCodonPos)) + codonGL = paramGL[pos_array] + + codonSeq = sapply(pos,function(x){ + seqP = paramGL[getCodonPos(x)] + muCodonPos = {x-1}%%3+1 + seqP[muCodonPos] = paramSeq[x] + return(seqP) + }) + GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) + Seqcodons = apply(codonSeq,2,c2s) + mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + names(mutationInfo) = mutationPos + } + if(any(!is.na(mutationInfo))){ + return(mutationInfo[!is.na(mutationInfo)]) + }else{ + return(NA) + } + + + }else{ + return (NA) + } + } + + processNucMutations2 <- function(mu){ + if(!is.na(mu)){ + #R + if(any(mu=="R")){ + Rs = mu[mu=="R"] + nucNumbs = as.numeric(names(Rs)) + R_CDR = sum(as.integer(CDR_Nuc[nucNumbs]),na.rm=T) + R_FWR = sum(as.integer(FWR_Nuc[nucNumbs]),na.rm=T) + }else{ + R_CDR = 0 + R_FWR = 0 + } + + #S + if(any(mu=="S")){ + Ss = mu[mu=="S"] + nucNumbs = as.numeric(names(Ss)) + S_CDR = sum(as.integer(CDR_Nuc[nucNumbs]),na.rm=T) + S_FWR = sum(as.integer(FWR_Nuc[nucNumbs]),na.rm=T) + }else{ + S_CDR = 0 + S_FWR = 0 + } + + + retVec = c(R_CDR,S_CDR,R_FWR,S_FWR) + retVec[is.na(retVec)]=0 + return(retVec) + }else{ + return(rep(0,4)) + } + } + + + ## Z-score Test + computeZScore <- function(mat, test="Focused"){ + matRes <- matrix(NA,ncol=2,nrow=(nrow(mat))) + if(test=="Focused"){ + #Z_Focused_CDR + #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) + P = apply(mat[,c(5,6,8)],1,function(x){(x[1]/sum(x))}) + R_mean = apply(cbind(mat[,c(1,2,4)],P),1,function(x){x[4]*(sum(x[1:3]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,1] = (mat[,1]-R_mean)/R_sd + + #Z_Focused_FWR + #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) + P = apply(mat[,c(7,6,8)],1,function(x){(x[1]/sum(x))}) + R_mean = apply(cbind(mat[,c(3,2,4)],P),1,function(x){x[4]*(sum(x[1:3]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,2] = (mat[,3]-R_mean)/R_sd + } + + if(test=="Local"){ + #Z_Focused_CDR + #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) + P = apply(mat[,c(5,6)],1,function(x){(x[1]/sum(x))}) + R_mean = apply(cbind(mat[,c(1,2)],P),1,function(x){x[3]*(sum(x[1:2]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,1] = (mat[,1]-R_mean)/R_sd + + #Z_Focused_FWR + #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) + P = apply(mat[,c(7,8)],1,function(x){(x[1]/sum(x))}) + R_mean = apply(cbind(mat[,c(3,4)],P),1,function(x){x[3]*(sum(x[1:2]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,2] = (mat[,3]-R_mean)/R_sd + } + + if(test=="Imbalanced"){ + #Z_Focused_CDR + #P_Denom = sum( mat[1,c(5,6,8)], na.rm=T ) + P = apply(mat[,5:8],1,function(x){((x[1]+x[2])/sum(x))}) + R_mean = apply(cbind(mat[,1:4],P),1,function(x){x[5]*(sum(x[1:4]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,1] = (mat[,1]-R_mean)/R_sd + + #Z_Focused_FWR + #P_Denom = sum( mat[1,c(7,6,8)], na.rm=T ) + P = apply(mat[,5:8],1,function(x){((x[3]+x[4])/sum(x))}) + R_mean = apply(cbind(mat[,1:4],P),1,function(x){x[5]*(sum(x[1:4]))}) + R_sd=sqrt(R_mean*(1-P)) + matRes[,2] = (mat[,3]-R_mean)/R_sd + } + + matRes[is.nan(matRes)] = NA + return(matRes) + } + + # Return a p-value for a z-score + z2p <- function(z){ + p=NA + if( !is.nan(z) && !is.na(z)){ + if(z>0){ + p = (1 - pnorm(z,0,1)) + } else if(z<0){ + p = (-1 * pnorm(z,0,1)) + } else{ + p = 0.5 + } + }else{ + p = NA + } + return(p) + } + + + ## Bayesian Test + + # Fitted parameter for the bayesian framework +BAYESIAN_FITTED<-c(0.407277142798302, 0.554007336744485, 0.63777155771234, 0.693989162719009, 0.735450014674917, 0.767972534429806, 0.794557287143399, 0.816906816601605, 0.83606796225341, 0.852729446430296, 0.867370424541641, 0.880339760590323, 0.891900995024999, 0.902259181289864, 0.911577919359,0.919990301665853, 0.927606458124537, 0.934518806350661, 0.940805863754375, 0.946534836475715, 0.951763691199255, 0.95654428191308, 0.960920179487397, 0.964930893680829, 0.968611312149038, 0.971992459313836, 0.975102110004818, 0.977964943023096, 0.980603428208439, 0.983037660179428, 0.985285800977406, 0.987364285326685, 0.989288037855441, 0.991070478823525, 0.992723699729969, 0.994259575477392, 0.995687688867975, 0.997017365051493, 0.998257085153047, 0.999414558305388, 1.00049681357804, 1.00151036237481, 1.00246080204981, 1.00335370751909, 1.0041939329768, 1.0049859393417, 1.00573382091263, 1.00644127217376, 1.00711179729107, 1.00774845526417, 1.00835412715854, 1.00893143010366, 1.00948275846309, 1.01001030293661, 1.01051606798079, 1.01100188771288, 1.01146944044216, 1.01192026195449, 1.01235575766094, 1.01277721370986) + CONST_i <- sort(c(((2^(seq(-39,0,length.out=201)))/2)[1:200],(c(0:11,13:99)+0.5)/100,1-(2^(seq(-39,0,length.out=201)))/2)) + + # Given x, M & p, returns a pdf + calculate_bayes <- function ( x=3, N=10, p=0.33, + i=CONST_i, + max_sigma=20,length_sigma=4001 + ){ + if(!0%in%N){ + G <- max(length(x),length(N),length(p)) + x=array(x,dim=G) + N=array(N,dim=G) + p=array(p,dim=G) + sigma_s<-seq(-max_sigma,max_sigma,length.out=length_sigma) + sigma_1<-log({i/{1-i}}/{p/{1-p}}) + index<-min(N,60) + y<-dbeta(i,x+BAYESIAN_FITTED[index],N+BAYESIAN_FITTED[index]-x)*(1-p)*p*exp(sigma_1)/({1-p}^2+2*p*{1-p}*exp(sigma_1)+{p^2}*exp(2*sigma_1)) + if(!sum(is.na(y))){ + tmp<-approx(sigma_1,y,sigma_s)$y + tmp/sum(tmp)/{2*max_sigma/{length_sigma-1}} + }else{ + return(NA) + } + }else{ + return(NA) + } + } + # Given a mat of observed & expected, return a list of CDR & FWR pdf for selection + computeBayesianScore <- function(mat, test="Focused", max_sigma=20,length_sigma=4001){ + flagOneSeq = F + if(nrow(mat)==1){ + mat=rbind(mat,mat) + flagOneSeq = T + } + if(test=="Focused"){ + #CDR + P = c(apply(mat[,c(5,6,8)],1,function(x){(x[1]/sum(x))}),0.5) + N = c(apply(mat[,c(1,2,4)],1,function(x){(sum(x))}),0) + X = c(mat[,1],0) + bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesCDR = bayesCDR[-length(bayesCDR)] + + #FWR + P = c(apply(mat[,c(7,6,8)],1,function(x){(x[1]/sum(x))}),0.5) + N = c(apply(mat[,c(3,2,4)],1,function(x){(sum(x))}),0) + X = c(mat[,3],0) + bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesFWR = bayesFWR[-length(bayesFWR)] + } + + if(test=="Local"){ + #CDR + P = c(apply(mat[,c(5,6)],1,function(x){(x[1]/sum(x))}),0.5) + N = c(apply(mat[,c(1,2)],1,function(x){(sum(x))}),0) + X = c(mat[,1],0) + bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesCDR = bayesCDR[-length(bayesCDR)] + + #FWR + P = c(apply(mat[,c(7,8)],1,function(x){(x[1]/sum(x))}),0.5) + N = c(apply(mat[,c(3,4)],1,function(x){(sum(x))}),0) + X = c(mat[,3],0) + bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesFWR = bayesFWR[-length(bayesFWR)] + } + + if(test=="Imbalanced"){ + #CDR + P = c(apply(mat[,c(5:8)],1,function(x){((x[1]+x[2])/sum(x))}),0.5) + N = c(apply(mat[,c(1:4)],1,function(x){(sum(x))}),0) + X = c(apply(mat[,c(1:2)],1,function(x){(sum(x))}),0) + bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesCDR = bayesCDR[-length(bayesCDR)] + + #FWR + P = c(apply(mat[,c(5:8)],1,function(x){((x[3]+x[4])/sum(x))}),0.5) + N = c(apply(mat[,c(1:4)],1,function(x){(sum(x))}),0) + X = c(apply(mat[,c(3:4)],1,function(x){(sum(x))}),0) + bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesFWR = bayesFWR[-length(bayesFWR)] + } + + if(test=="ImbalancedSilent"){ + #CDR + P = c(apply(mat[,c(6,8)],1,function(x){((x[1])/sum(x))}),0.5) + N = c(apply(mat[,c(2,4)],1,function(x){(sum(x))}),0) + X = c(apply(mat[,c(2,4)],1,function(x){(x[1])}),0) + bayesCDR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesCDR = bayesCDR[-length(bayesCDR)] + + #FWR + P = c(apply(mat[,c(6,8)],1,function(x){((x[2])/sum(x))}),0.5) + N = c(apply(mat[,c(2,4)],1,function(x){(sum(x))}),0) + X = c(apply(mat[,c(2,4)],1,function(x){(x[2])}),0) + bayesFWR = apply(cbind(X,N,P),1,function(x){calculate_bayes(x=x[1],N=x[2],p=x[3],max_sigma=max_sigma,length_sigma=length_sigma)}) + bayesFWR = bayesFWR[-length(bayesFWR)] + } + + if(flagOneSeq==T){ + bayesCDR = bayesCDR[1] + bayesFWR = bayesFWR[1] + } + return( list("CDR"=bayesCDR, "FWR"=bayesFWR) ) + } + + ##Covolution + break2chunks<-function(G=1000){ + base<-2^round(log(sqrt(G),2),0) + return(c(rep(base,floor(G/base)-1),base+G-(floor(G/base)*base))) + } + + PowersOfTwo <- function(G=100){ + exponents <- array() + i = 0 + while(G > 0){ + i=i+1 + exponents[i] <- floor( log2(G) ) + G <- G-2^exponents[i] + } + return(exponents) + } + + convolutionPowersOfTwo <- function( cons, length_sigma=4001 ){ + G = ncol(cons) + if(G>1){ + for(gen in log(G,2):1){ + ll<-seq(from=2,to=2^gen,by=2) + sapply(ll,function(l){cons[,l/2]<<-weighted_conv(cons[,l],cons[,l-1],length_sigma=length_sigma)}) + } + } + return( cons[,1] ) + } + + convolutionPowersOfTwoByTwos <- function( cons, length_sigma=4001,G=1 ){ + if(length(ncol(cons))) G<-ncol(cons) + groups <- PowersOfTwo(G) + matG <- matrix(NA, ncol=length(groups), nrow=length(cons)/G ) + startIndex = 1 + for( i in 1:length(groups) ){ + stopIndex <- 2^groups[i] + startIndex - 1 + if(stopIndex!=startIndex){ + matG[,i] <- convolutionPowersOfTwo( cons[,startIndex:stopIndex], length_sigma=length_sigma ) + startIndex = stopIndex + 1 + } + else { + if(G>1) matG[,i] <- cons[,startIndex:stopIndex] + else matG[,i] <- cons + #startIndex = stopIndex + 1 + } + } + return( list( matG, groups ) ) + } + + weighted_conv<-function(x,y,w=1,m=100,length_sigma=4001){ + lx<-length(x) + ly<-length(y) + if({lx1){ + while( i1 & Length_Postrior<=Threshold){ + cons = matrix(unlist(listPosteriors),length(listPosteriors[[1]]),length(listPosteriors)) + listMatG <- convolutionPowersOfTwoByTwos(cons,length_sigma=length_sigma) + y<-calculate_bayesGHelper(listMatG,length_sigma=length_sigma) + return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) + }else if(Length_Postrior==1) return(listPosteriors[[1]]) + else if(Length_Postrior==0) return(NA) + else { + cons = matrix(unlist(listPosteriors),length(listPosteriors[[1]]),length(listPosteriors)) + y = fastConv(cons,max_sigma=max_sigma, length_sigma=length_sigma ) + return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) + } + } + + fastConv<-function(cons, max_sigma=20, length_sigma=4001){ + chunks<-break2chunks(G=ncol(cons)) + if(ncol(cons)==3) chunks<-2:1 + index_chunks_end <- cumsum(chunks) + index_chunks_start <- c(1,index_chunks_end[-length(index_chunks_end)]+1) + index_chunks <- cbind(index_chunks_start,index_chunks_end) + + case <- sum(chunks!=chunks[1]) + if(case==1) End <- max(1,((length(index_chunks)/2)-1)) + else End <- max(1,((length(index_chunks)/2))) + + firsts <- sapply(1:End,function(i){ + indexes<-index_chunks[i,1]:index_chunks[i,2] + convolutionPowersOfTwoByTwos(cons[ ,indexes])[[1]] + }) + if(case==0){ + result<-calculate_bayesGHelper( convolutionPowersOfTwoByTwos(firsts) ) + }else if(case==1){ + last<-list(calculate_bayesGHelper( + convolutionPowersOfTwoByTwos( cons[ ,index_chunks[length(index_chunks)/2,1]:index_chunks[length(index_chunks)/2,2]] ) + ),0) + result_first<-calculate_bayesGHelper(convolutionPowersOfTwoByTwos(firsts)) + result<-calculate_bayesGHelper( + list( + cbind( + result_first,last[[1]]), + c(log(index_chunks_end[length(index_chunks)/2-1],2),log(index_chunks[length(index_chunks)/2,2]-index_chunks[length(index_chunks)/2,1]+1,2)) + ) + ) + } + return(as.vector(result)) + } + + # Computes the 95% CI for a pdf + calcBayesCI <- function(Pdf,low=0.025,up=0.975,max_sigma=20, length_sigma=4001){ + if(length(Pdf)!=length_sigma) return(NA) + sigma_s=seq(-max_sigma,max_sigma,length.out=length_sigma) + cdf = cumsum(Pdf) + cdf = cdf/cdf[length(cdf)] + return( c(sigma_s[findInterval(low,cdf)-1] , sigma_s[findInterval(up,cdf)]) ) + } + + # Computes a mean for a pdf + calcBayesMean <- function(Pdf,max_sigma=20,length_sigma=4001){ + if(length(Pdf)!=length_sigma) return(NA) + sigma_s=seq(-max_sigma,max_sigma,length.out=length_sigma) + norm = {length_sigma-1}/2/max_sigma + return( (Pdf%*%sigma_s/norm) ) + } + + # Returns the mean, and the 95% CI for a pdf + calcBayesOutputInfo <- function(Pdf,low=0.025,up=0.975,max_sigma=20, length_sigma=4001){ + if(is.na(Pdf)) + return(rep(NA,3)) + bCI = calcBayesCI(Pdf=Pdf,low=low,up=up,max_sigma=max_sigma,length_sigma=length_sigma) + bMean = calcBayesMean(Pdf=Pdf,max_sigma=max_sigma,length_sigma=length_sigma) + return(c(bMean, bCI)) + } + + # Computes the p-value of a pdf + computeSigmaP <- function(Pdf, length_sigma=4001, max_sigma=20){ + if(length(Pdf)>1){ + norm = {length_sigma-1}/2/max_sigma + pVal = {sum(Pdf[1:{{length_sigma-1}/2}]) + Pdf[{{length_sigma+1}/2}]/2}/norm + if(pVal>0.5){ + pVal = pVal-1 + } + return(pVal) + }else{ + return(NA) + } + } + + # Compute p-value of two distributions + compareTwoDistsFaster <-function(sigma_S=seq(-20,20,length.out=4001), N=10000, dens1=runif(4001,0,1), dens2=runif(4001,0,1)){ + #print(c(length(dens1),length(dens2))) + if(length(dens1)>1 & length(dens2)>1 ){ + dens1<-dens1/sum(dens1) + dens2<-dens2/sum(dens2) + cum2 <- cumsum(dens2)-dens2/2 + tmp<- sum(sapply(1:length(dens1),function(i)return(dens1[i]*cum2[i]))) + #print(tmp) + if(tmp>0.5)tmp<-tmp-1 + return( tmp ) + } + else { + return(NA) + } + #return (sum(sapply(1:N,function(i)(sample(sigma_S,1,prob=dens1)>sample(sigma_S,1,prob=dens2))))/N) + } + + # get number of seqeunces contributing to the sigma (i.e. seqeunces with mutations) + numberOfSeqsWithMutations <- function(matMutations,test=1){ + if(test==4)test=2 + cdrSeqs <- 0 + fwrSeqs <- 0 + if(test==1){#focused + cdrMutations <- apply(matMutations, 1, function(x){ sum(x[c(1,2,4)]) }) + fwrMutations <- apply(matMutations, 1, function(x){ sum(x[c(3,4,2)]) }) + if( any(which(cdrMutations>0)) ) cdrSeqs <- sum(cdrMutations>0) + if( any(which(fwrMutations>0)) ) fwrSeqs <- sum(fwrMutations>0) + } + if(test==2){#local + cdrMutations <- apply(matMutations, 1, function(x){ sum(x[c(1,2)]) }) + fwrMutations <- apply(matMutations, 1, function(x){ sum(x[c(3,4)]) }) + if( any(which(cdrMutations>0)) ) cdrSeqs <- sum(cdrMutations>0) + if( any(which(fwrMutations>0)) ) fwrSeqs <- sum(fwrMutations>0) + } + return(c("CDR"=cdrSeqs, "FWR"=fwrSeqs)) +} + + + +shadeColor <- function(sigmaVal=NA,pVal=NA){ + if(is.na(sigmaVal) & is.na(pVal)) return(NA) + if(is.na(sigmaVal) & !is.na(pVal)) sigmaVal=sign(pVal) + if(is.na(pVal) || pVal==1 || pVal==0){ + returnColor = "#FFFFFF"; + }else{ + colVal=abs(pVal); + + if(sigmaVal<0){ + if(colVal>0.1) + returnColor = "#CCFFCC"; + if(colVal<=0.1) + returnColor = "#99FF99"; + if(colVal<=0.050) + returnColor = "#66FF66"; + if(colVal<=0.010) + returnColor = "#33FF33"; + if(colVal<=0.005) + returnColor = "#00FF00"; + + }else{ + if(colVal>0.1) + returnColor = "#FFCCCC"; + if(colVal<=0.1) + returnColor = "#FF9999"; + if(colVal<=0.05) + returnColor = "#FF6666"; + if(colVal<=0.01) + returnColor = "#FF3333"; + if(colVal<0.005) + returnColor = "#FF0000"; + } + } + + return(returnColor) +} + + + +plotHelp <- function(xfrac=0.05,yfrac=0.05,log=FALSE){ + if(!log){ + x = par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac + y = par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac + }else { + if(log==2){ + x = par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac + y = 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac) + } + if(log==1){ + x = 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac) + y = par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac + } + if(log==3){ + x = 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac) + y = 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac) + } + } + return(c("x"=x,"y"=y)) +} + +# SHMulation + + # Based on targeting, introduce a single mutation & then update the targeting + oneMutation <- function(){ + # Pick a postion + mutation + posMutation = sample(1:(seqGermlineLen*4),1,replace=F,prob=as.vector(seqTargeting)) + posNucNumb = ceiling(posMutation/4) # Nucleotide number + posNucKind = 4 - ( (posNucNumb*4) - posMutation ) # Nuc the position mutates to + + #mutate the simulation sequence + seqSimVec <- s2c(seqSim) + seqSimVec[posNucNumb] <- NUCLEOTIDES[posNucKind] + seqSim <<- c2s(seqSimVec) + + #update Mutability, Targeting & MutationsTypes + updateMutabilityNTargeting(posNucNumb) + + #return(c(posNucNumb,NUCLEOTIDES[posNucKind])) + return(posNucNumb) + } + + updateMutabilityNTargeting <- function(position){ + min_i<-max((position-2),1) + max_i<-min((position+2),nchar(seqSim)) + min_ii<-min(min_i,3) + + #mutability - update locally + seqMutability[(min_i):(max_i)] <<- computeMutabilities(substr(seqSim,position-4,position+4))[(min_ii):(max_i-min_i+min_ii)] + + + #targeting - compute locally + seqTargeting[,min_i:max_i] <<- computeTargeting(substr(seqSim,min_i,max_i),seqMutability[min_i:max_i]) + seqTargeting[is.na(seqTargeting)] <<- 0 + #mutCodonPos = getCodonPos(position) + mutCodonPos = seq(getCodonPos(min_i)[1],getCodonPos(max_i)[3]) + #cat(mutCodonPos,"\n") + mutTypeCodon = getCodonPos(position) + seqMutationTypes[,mutTypeCodon] <<- computeMutationTypesFast( substr(seqSim,mutTypeCodon[1],mutTypeCodon[3]) ) + # Stop = 0 + if(any(seqMutationTypes[,mutCodonPos]=="Stop",na.rm=T )){ + seqTargeting[,mutCodonPos][seqMutationTypes[,mutCodonPos]=="Stop"] <<- 0 + } + + + #Selection + selectedPos = (min_i*4-4)+(which(seqMutationTypes[,min_i:max_i]=="R")) + # CDR + selectedCDR = selectedPos[which(matCDR[selectedPos]==T)] + seqTargeting[selectedCDR] <<- seqTargeting[selectedCDR] * exp(selCDR) + seqTargeting[selectedCDR] <<- seqTargeting[selectedCDR]/baseLineCDR_K + + # FWR + selectedFWR = selectedPos[which(matFWR[selectedPos]==T)] + seqTargeting[selectedFWR] <<- seqTargeting[selectedFWR] * exp(selFWR) + seqTargeting[selectedFWR] <<- seqTargeting[selectedFWR]/baseLineFWR_K + + } + + + + # Validate the mutation: if the mutation has not been sampled before validate it, else discard it. + validateMutation <- function(){ + if( !(mutatedPos%in%mutatedPositions) ){ # if it's a new mutation + uniqueMutationsIntroduced <<- uniqueMutationsIntroduced + 1 + mutatedPositions[uniqueMutationsIntroduced] <<- mutatedPos + }else{ + if(substr(seqSim,mutatedPos,mutatedPos)==substr(seqGermline,mutatedPos,mutatedPos)){ # back to germline mutation + mutatedPositions <<- mutatedPositions[-which(mutatedPositions==mutatedPos)] + uniqueMutationsIntroduced <<- uniqueMutationsIntroduced - 1 + } + } + } + + + + # Places text (labels) at normalized coordinates + myaxis <- function(xfrac=0.05,yfrac=0.05,log=FALSE,w="text",cex=1,adj=1,thecol="black"){ + par(xpd=TRUE) + if(!log) + text(par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac,par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac,w,cex=cex,adj=adj,col=thecol) + else { + if(log==2) + text( + par()$usr[1]-(par()$usr[2]-par()$usr[1])*xfrac, + 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac), + w,cex=cex,adj=adj,col=thecol) + if(log==1) + text( + 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac), + par()$usr[4]+(par()$usr[4]-par()$usr[3])*yfrac, + w,cex=cex,adj=adj,col=thecol) + if(log==3) + text( + 10^((par()$usr[1])-((par()$usr[2])-(par()$usr[1]))*xfrac), + 10^((par()$usr[4])+((par()$usr[4])-(par()$usr[3]))*yfrac), + w,cex=cex,adj=adj,col=thecol) + } + par(xpd=FALSE) + } + + + + # Count the mutations in a sequence + analyzeMutations <- function( inputMatrixIndex, model = 0 , multipleMutation=0, seqWithStops=0){ + + paramGL = s2c(matInput[inputMatrixIndex,2]) + paramSeq = s2c(matInput[inputMatrixIndex,1]) + + #if( any(paramSeq=="N") ){ + # gapPos_Seq = which(paramSeq=="N") + # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] + # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] + #} + mutations_val = paramGL != paramSeq + + if(any(mutations_val)){ + mutationPos = which(mutations_val)#{1:length(mutations_val)}[mutations_val] + length_mutations =length(mutationPos) + mutationInfo = rep(NA,length_mutations) + + pos<- mutationPos + pos_array<-array(sapply(pos,getCodonPos)) + codonGL = paramGL[pos_array] + codonSeqWhole = paramSeq[pos_array] + codonSeq = sapply(pos,function(x){ + seqP = paramGL[getCodonPos(x)] + muCodonPos = {x-1}%%3+1 + seqP[muCodonPos] = paramSeq[x] + return(seqP) + }) + GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) + SeqcodonsWhole = apply(matrix(codonSeqWhole,length_mutations,3,byrow=TRUE),1,c2s) + Seqcodons = apply(codonSeq,2,c2s) + + mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + names(mutationInfo) = mutationPos + + mutationInfoWhole = apply(rbind(GLcodons , SeqcodonsWhole),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + names(mutationInfoWhole) = mutationPos + + mutationInfo <- mutationInfo[!is.na(mutationInfo)] + mutationInfoWhole <- mutationInfoWhole[!is.na(mutationInfoWhole)] + + if(any(!is.na(mutationInfo))){ + + #Filter based on Stop (at the codon level) + if(seqWithStops==1){ + nucleotidesAtStopCodons = names(mutationInfoWhole[mutationInfoWhole!="Stop"]) + mutationInfo = mutationInfo[nucleotidesAtStopCodons] + mutationInfoWhole = mutationInfo[nucleotidesAtStopCodons] + }else{ + countStops = sum(mutationInfoWhole=="Stop") + if(seqWithStops==2 & countStops==0) mutationInfo = NA + if(seqWithStops==3 & countStops>0) mutationInfo = NA + } + + if(any(!is.na(mutationInfo))){ + #Filter mutations based on multipleMutation + if(multipleMutation==1 & !is.na(mutationInfo)){ + mutationCodons = getCodonNumb(as.numeric(names(mutationInfoWhole))) + tableMutationCodons <- table(mutationCodons) + codonsWithMultipleMutations <- as.numeric(names(tableMutationCodons[tableMutationCodons>1])) + if(any(codonsWithMultipleMutations)){ + #remove the nucleotide mutations in the codons with multiple mutations + mutationInfo <- mutationInfo[!(mutationCodons %in% codonsWithMultipleMutations)] + #replace those codons with Ns in the input sequence + paramSeq[unlist(lapply(codonsWithMultipleMutations, getCodonNucs))] = "N" + matInput[inputMatrixIndex,1] <<- c2s(paramSeq) + } + } + + #Filter mutations based on the model + if(any(mutationInfo)==T | is.na(any(mutationInfo))){ + + if(model==1 & !is.na(mutationInfo)){ + mutationInfo <- mutationInfo[mutationInfo=="S"] + } + if(any(mutationInfo)==T | is.na(any(mutationInfo))) return(mutationInfo) + else return(NA) + }else{ + return(NA) + } + }else{ + return(NA) + } + + + }else{ + return(NA) + } + + + }else{ + return (NA) + } + } + + analyzeMutationsFixed <- function( inputArray, model = 0 , multipleMutation=0, seqWithStops=0){ + + paramGL = s2c(inputArray[2]) + paramSeq = s2c(inputArray[1]) + inputSeq <- inputArray[1] + #if( any(paramSeq=="N") ){ + # gapPos_Seq = which(paramSeq=="N") + # gapPos_Seq_ToReplace = gapPos_Seq[paramGL[gapPos_Seq] != "N"] + # paramSeq[gapPos_Seq_ToReplace] = paramGL[gapPos_Seq_ToReplace] + #} + mutations_val = paramGL != paramSeq + + if(any(mutations_val)){ + mutationPos = which(mutations_val)#{1:length(mutations_val)}[mutations_val] + length_mutations =length(mutationPos) + mutationInfo = rep(NA,length_mutations) + + pos<- mutationPos + pos_array<-array(sapply(pos,getCodonPos)) + codonGL = paramGL[pos_array] + codonSeqWhole = paramSeq[pos_array] + codonSeq = sapply(pos,function(x){ + seqP = paramGL[getCodonPos(x)] + muCodonPos = {x-1}%%3+1 + seqP[muCodonPos] = paramSeq[x] + return(seqP) + }) + GLcodons = apply(matrix(codonGL,length_mutations,3,byrow=TRUE),1,c2s) + SeqcodonsWhole = apply(matrix(codonSeqWhole,length_mutations,3,byrow=TRUE),1,c2s) + Seqcodons = apply(codonSeq,2,c2s) + + mutationInfo = apply(rbind(GLcodons , Seqcodons),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + names(mutationInfo) = mutationPos + + mutationInfoWhole = apply(rbind(GLcodons , SeqcodonsWhole),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + names(mutationInfoWhole) = mutationPos + + mutationInfo <- mutationInfo[!is.na(mutationInfo)] + mutationInfoWhole <- mutationInfoWhole[!is.na(mutationInfoWhole)] + + if(any(!is.na(mutationInfo))){ + + #Filter based on Stop (at the codon level) + if(seqWithStops==1){ + nucleotidesAtStopCodons = names(mutationInfoWhole[mutationInfoWhole!="Stop"]) + mutationInfo = mutationInfo[nucleotidesAtStopCodons] + mutationInfoWhole = mutationInfo[nucleotidesAtStopCodons] + }else{ + countStops = sum(mutationInfoWhole=="Stop") + if(seqWithStops==2 & countStops==0) mutationInfo = NA + if(seqWithStops==3 & countStops>0) mutationInfo = NA + } + + if(any(!is.na(mutationInfo))){ + #Filter mutations based on multipleMutation + if(multipleMutation==1 & !is.na(mutationInfo)){ + mutationCodons = getCodonNumb(as.numeric(names(mutationInfoWhole))) + tableMutationCodons <- table(mutationCodons) + codonsWithMultipleMutations <- as.numeric(names(tableMutationCodons[tableMutationCodons>1])) + if(any(codonsWithMultipleMutations)){ + #remove the nucleotide mutations in the codons with multiple mutations + mutationInfo <- mutationInfo[!(mutationCodons %in% codonsWithMultipleMutations)] + #replace those codons with Ns in the input sequence + paramSeq[unlist(lapply(codonsWithMultipleMutations, getCodonNucs))] = "N" + #matInput[inputMatrixIndex,1] <<- c2s(paramSeq) + inputSeq <- c2s(paramSeq) + } + } + + #Filter mutations based on the model + if(any(mutationInfo)==T | is.na(any(mutationInfo))){ + + if(model==1 & !is.na(mutationInfo)){ + mutationInfo <- mutationInfo[mutationInfo=="S"] + } + if(any(mutationInfo)==T | is.na(any(mutationInfo))) return(list(mutationInfo,inputSeq)) + else return(list(NA,inputSeq)) + }else{ + return(list(NA,inputSeq)) + } + }else{ + return(list(NA,inputSeq)) + } + + + }else{ + return(list(NA,inputSeq)) + } + + + }else{ + return (list(NA,inputSeq)) + } + } + + # triMutability Background Count + buildMutabilityModel <- function( inputMatrixIndex, model=0 , multipleMutation=0, seqWithStops=0, stopMutations=0){ + + #rowOrigMatInput = matInput[inputMatrixIndex,] + seqGL = gsub("-", "", matInput[inputMatrixIndex,2]) + seqInput = gsub("-", "", matInput[inputMatrixIndex,1]) + #matInput[inputMatrixIndex,] <<- cbind(seqInput,seqGL) + tempInput <- cbind(seqInput,seqGL) + seqLength = nchar(seqGL) + list_analyzeMutationsFixed<- analyzeMutationsFixed(tempInput, model, multipleMutation, seqWithStops) + mutationCount <- list_analyzeMutationsFixed[[1]] + seqInput <- list_analyzeMutationsFixed[[2]] + BackgroundMatrix = mutabilityMatrix + MutationMatrix = mutabilityMatrix + MutationCountMatrix = mutabilityMatrix + if(!is.na(mutationCount)){ + if((stopMutations==0 & model==0) | (stopMutations==1 & (sum(mutationCount=="Stop")0)) ){ + + fivermerStartPos = 1:(seqLength-4) + fivemerLength <- length(fivermerStartPos) + fivemerGL <- substr(rep(seqGL,length(fivermerStartPos)),(fivermerStartPos),(fivermerStartPos+4)) + fivemerSeq <- substr(rep(seqInput,length(fivermerStartPos)),(fivermerStartPos),(fivermerStartPos+4)) + + #Background + for(fivemerIndex in 1:fivemerLength){ + fivemer = fivemerGL[fivemerIndex] + if(!any(grep("N",fivemer))){ + fivemerCodonPos = fivemerCodon(fivemerIndex) + fivemerReadingFrameCodon = substr(fivemer,fivemerCodonPos[1],fivemerCodonPos[3]) + fivemerReadingFrameCodonInputSeq = substr(fivemerSeq[fivemerIndex],fivemerCodonPos[1],fivemerCodonPos[3]) + + # All mutations model + #if(!any(grep("N",fivemerReadingFrameCodon))){ + if(model==0){ + if(stopMutations==0){ + if(!any(grep("N",fivemerReadingFrameCodonInputSeq))) + BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + 1) + }else{ + if( !any(grep("N",fivemerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(fivemerReadingFrameCodon)!="*" ){ + positionWithinCodon = which(fivemerCodonPos==3)#positionsWithinCodon[(fivemerCodonPos[1]%%3)+1] + BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + probNonStopMutations[fivemerReadingFrameCodon,positionWithinCodon]) + } + } + }else{ # Only silent mutations + if( !any(grep("N",fivemerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(fivemerReadingFrameCodon)!="*" & translateCodonToAminoAcid(fivemerReadingFrameCodonInputSeq)==translateCodonToAminoAcid(fivemerReadingFrameCodon)){ + positionWithinCodon = which(fivemerCodonPos==3) + BackgroundMatrix[fivemer] <- (BackgroundMatrix[fivemer] + probSMutations[fivemerReadingFrameCodon,positionWithinCodon]) + } + } + #} + } + } + + #Mutations + if(stopMutations==1) mutationCount = mutationCount[mutationCount!="Stop"] + if(model==1) mutationCount = mutationCount[mutationCount=="S"] + mutationPositions = as.numeric(names(mutationCount)) + mutationCount = mutationCount[mutationPositions>2 & mutationPositions<(seqLength-1)] + mutationPositions = mutationPositions[mutationPositions>2 & mutationPositions<(seqLength-1)] + countMutations = 0 + for(mutationPosition in mutationPositions){ + fivemerIndex = mutationPosition-2 + fivemer = fivemerSeq[fivemerIndex] + GLfivemer = fivemerGL[fivemerIndex] + fivemerCodonPos = fivemerCodon(fivemerIndex) + fivemerReadingFrameCodon = substr(fivemer,fivemerCodonPos[1],fivemerCodonPos[3]) + fivemerReadingFrameCodonGL = substr(GLfivemer,fivemerCodonPos[1],fivemerCodonPos[3]) + if(!any(grep("N",fivemer)) & !any(grep("N",GLfivemer))){ + if(model==0){ + countMutations = countMutations + 1 + MutationMatrix[GLfivemer] <- (MutationMatrix[GLfivemer] + 1) + MutationCountMatrix[GLfivemer] <- (MutationCountMatrix[GLfivemer] + 1) + }else{ + if( translateCodonToAminoAcid(fivemerReadingFrameCodonGL)!="*" ){ + countMutations = countMutations + 1 + positionWithinCodon = which(fivemerCodonPos==3) + glNuc = substr(fivemerReadingFrameCodonGL,positionWithinCodon,positionWithinCodon) + inputNuc = substr(fivemerReadingFrameCodon,positionWithinCodon,positionWithinCodon) + MutationMatrix[GLfivemer] <- (MutationMatrix[GLfivemer] + substitution[glNuc,inputNuc]) + MutationCountMatrix[GLfivemer] <- (MutationCountMatrix[GLfivemer] + 1) + } + } + } + } + + seqMutability = MutationMatrix/BackgroundMatrix + seqMutability = seqMutability/sum(seqMutability,na.rm=TRUE) + #cat(inputMatrixIndex,"\t",countMutations,"\n") + return(list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix, "BackgroundMatrix"=BackgroundMatrix)) + + } + } + + } + + #Returns the codon position containing the middle nucleotide + fivemerCodon <- function(fivemerIndex){ + codonPos = list(2:4,1:3,3:5) + fivemerType = fivemerIndex%%3 + return(codonPos[[fivemerType+1]]) + } + + #returns probability values for one mutation in codons resulting in R, S or Stop + probMutations <- function(typeOfMutation){ + matMutationProb <- matrix(0,ncol=3,nrow=125,dimnames=list(words(alphabet = c(NUCLEOTIDES,"N"), length=3),c(1:3))) + for(codon in rownames(matMutationProb)){ + if( !any(grep("N",codon)) ){ + for(muPos in 1:3){ + matCodon = matrix(rep(s2c(codon),3),nrow=3,ncol=3,byrow=T) + glNuc = matCodon[1,muPos] + matCodon[,muPos] = canMutateTo(glNuc) + substitutionRate = substitution[glNuc,matCodon[,muPos]] + typeOfMutations = apply(rbind(rep(codon,3),apply(matCodon,1,c2s)),2,function(x){mutationType(c2s(x[1]),c2s(x[2]))}) + matMutationProb[codon,muPos] <- sum(substitutionRate[typeOfMutations==typeOfMutation]) + } + } + } + + return(matMutationProb) + } + + + + +#Mapping Trinucleotides to fivemers +mapTriToFivemer <- function(triMutability=triMutability_Literature_Human){ + rownames(triMutability) <- triMutability_Names + Fivemer<-rep(NA,1024) + names(Fivemer)<-words(alphabet=NUCLEOTIDES,length=5) + Fivemer<-sapply(names(Fivemer),function(Word)return(sum( c(triMutability[substring(Word,3,5),1],triMutability[substring(Word,2,4),2],triMutability[substring(Word,1,3),3]),na.rm=TRUE))) + Fivemer<-Fivemer/sum(Fivemer) + return(Fivemer) +} + +collapseFivemerToTri<-function(Fivemer,Weights=MutabilityWeights,position=1,NUC="A"){ + Indices<-substring(names(Fivemer),3,3)==NUC + Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) + tapply(which(Indices),Factors,function(i)weighted.mean(Fivemer[i],Weights[i],na.rm=TRUE)) +} + + + +CountFivemerToTri<-function(Fivemer,Weights=MutabilityWeights,position=1,NUC="A"){ + Indices<-substring(names(Fivemer),3,3)==NUC + Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) + tapply(which(Indices),Factors,function(i)sum(Weights[i],na.rm=TRUE)) +} + +#Uses the real counts of the mutated fivemers +CountFivemerToTri2<-function(Fivemer,Counts=MutabilityCounts,position=1,NUC="A"){ + Indices<-substring(names(Fivemer),3,3)==NUC + Factors<-substring(names(Fivemer[Indices]),(4-position),(6-position)) + tapply(which(Indices),Factors,function(i)sum(Counts[i],na.rm=TRUE)) +} + +bootstrap<-function(x=c(33,12,21),M=10000,alpha=0.05){ +N<-sum(x) +if(N){ +p<-x/N +k<-length(x)-1 +tmp<-rmultinom(M, size = N, prob=p) +tmp_p<-apply(tmp,2,function(y)y/N) +(apply(tmp_p,1,function(y)quantile(y,c(alpha/2/k,1-alpha/2/k)))) +} +else return(matrix(0,2,length(x))) +} + + + + +bootstrap2<-function(x=c(33,12,21),n=10,M=10000,alpha=0.05){ + +N<-sum(x) +k<-length(x) +y<-rep(1:k,x) +tmp<-sapply(1:M,function(i)sample(y,n)) +if(n>1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[,j]==i)))/n +if(n==1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[j]==i)))/n +(apply(tmp_p,1,function(z)quantile(z,c(alpha/2/(k-1),1-alpha/2/(k-1))))) +} + + + +p_value<-function(x=c(33,12,21),M=100000,x_obs=c(2,5,3)){ +n=sum(x_obs) +N<-sum(x) +k<-length(x) +y<-rep(1:k,x) +tmp<-sapply(1:M,function(i)sample(y,n)) +if(n>1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[,j]==i))) +if(n==1)tmp_p<-sapply(1:M,function(j)sapply(1:k,function(i)sum(tmp[j]==i))) +tmp<-rbind(sapply(1:3,function(i)sum(tmp_p[i,]>=x_obs[i])/M), +sapply(1:3,function(i)sum(tmp_p[i,]<=x_obs[i])/M)) +sapply(1:3,function(i){if(tmp[1,i]>=tmp[2,i])return(-tmp[2,i])else return(tmp[1,i])}) +} + +#"D:\\Sequences\\IMGT Germlines\\Human_SNPless_IGHJ.FASTA" +# Remove SNPs from IMGT germline segment alleles +generateUnambiguousRepertoire <- function(repertoireInFile,repertoireOutFile){ + repertoireIn <- read.fasta(repertoireInFile, seqtype="DNA",as.string=T,set.attributes=F,forceDNAtolower=F) + alleleNames <- sapply(names(repertoireIn),function(x)strsplit(x,"|",fixed=TRUE)[[1]][2]) + SNPs <- tapply(repertoireIn,sapply(alleleNames,function(x)strsplit(x,"*",fixed=TRUE)[[1]][1]),function(x){ + Indices<-NULL + for(i in 1:length(x)){ + firstSeq = s2c(x[[1]]) + iSeq = s2c(x[[i]]) + Indices<-c(Indices,which(firstSeq[1:320]!=iSeq[1:320] & firstSeq[1:320]!="." & iSeq[1:320]!="." )) + } + return(sort(unique(Indices))) + }) + repertoireOut <- repertoireIn + repertoireOut <- lapply(names(repertoireOut), function(repertoireName){ + alleleName <- strsplit(repertoireName,"|",fixed=TRUE)[[1]][2] + geneSegmentName <- strsplit(alleleName,"*",fixed=TRUE)[[1]][1] + alleleSeq <- s2c(repertoireOut[[repertoireName]]) + alleleSeq[as.numeric(unlist(SNPs[geneSegmentName]))] <- "N" + alleleSeq <- c2s(alleleSeq) + repertoireOut[[repertoireName]] <- alleleSeq + }) + names(repertoireOut) <- names(repertoireIn) + write.fasta(repertoireOut,names(repertoireOut),file.out=repertoireOutFile) + +} + + + + + + +############ +groupBayes2 = function(indexes, param_resultMat){ + + BayesGDist_Focused_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2]+x[4])})) + BayesGDist_Focused_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(3,2,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[3]+x[2]+x[4])})) + #BayesGDist_Local_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2])})) + #BayesGDist_Local_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[3]+x[4])})) + #BayesGDist_Global_CDR = calculate_bayesG( x=param_resultMat[indexes,1], N=apply(param_resultMat[indexes,c(1,2,3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[1]/(x[1]+x[2]+x[3]+x[4])})) + #BayesGDist_Global_FWR = calculate_bayesG( x=param_resultMat[indexes,3], N=apply(param_resultMat[indexes,c(1,2,3,4)],1,sum,na.rm=T), p=apply(param_resultMat[indexes,5:8],1,function(x){x[3]/(x[1]+x[2]+x[3]+x[4])})) + return ( list("BayesGDist_Focused_CDR"=BayesGDist_Focused_CDR, + "BayesGDist_Focused_FWR"=BayesGDist_Focused_FWR) ) + #"BayesGDist_Local_CDR"=BayesGDist_Local_CDR, + #"BayesGDist_Local_FWR" = BayesGDist_Local_FWR)) +# "BayesGDist_Global_CDR" = BayesGDist_Global_CDR, +# "BayesGDist_Global_FWR" = BayesGDist_Global_FWR) ) + + +} + + +calculate_bayesG <- function( x=array(), N=array(), p=array(), max_sigma=20, length_sigma=4001){ + G <- max(length(x),length(N),length(p)) + x=array(x,dim=G) + N=array(N,dim=G) + p=array(p,dim=G) + + indexOfZero = N>0 & p>0 + N = N[indexOfZero] + x = x[indexOfZero] + p = p[indexOfZero] + G <- length(x) + + if(G){ + + cons<-array( dim=c(length_sigma,G) ) + if(G==1) { + return(calculate_bayes(x=x[G],N=N[G],p=p[G],max_sigma=max_sigma,length_sigma=length_sigma)) + } + else { + for(g in 1:G) cons[,g] <- calculate_bayes(x=x[g],N=N[g],p=p[g],max_sigma=max_sigma,length_sigma=length_sigma) + listMatG <- convolutionPowersOfTwoByTwos(cons,length_sigma=length_sigma) + y<-calculate_bayesGHelper(listMatG,length_sigma=length_sigma) + return( y/sum(y)/(2*max_sigma/(length_sigma-1)) ) + } + }else{ + return(NA) + } +} + + +calculate_bayesGHelper <- function( listMatG,length_sigma=4001 ){ + matG <- listMatG[[1]] + groups <- listMatG[[2]] + i = 1 + resConv <- matG[,i] + denom <- 2^groups[i] + if(length(groups)>1){ + while( i0)) ){ + +# ONEmerStartPos = 1:(seqLength) +# ONEmerLength <- length(ONEmerStartPos) + ONEmerGL <- s2c(seqGL) + ONEmerSeq <- s2c(seqInput) + + #Background + for(ONEmerIndex in 1:seqLength){ + ONEmer = ONEmerGL[ONEmerIndex] + if(ONEmer!="N"){ + ONEmerCodonPos = getCodonPos(ONEmerIndex) + ONEmerReadingFrameCodon = c2s(ONEmerGL[ONEmerCodonPos]) + ONEmerReadingFrameCodonInputSeq = c2s(ONEmerSeq[ONEmerCodonPos] ) + + # All mutations model + #if(!any(grep("N",ONEmerReadingFrameCodon))){ + if(model==0){ + if(stopMutations==0){ + if(!any(grep("N",ONEmerReadingFrameCodonInputSeq))) + BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + 1) + }else{ + if( !any(grep("N",ONEmerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)!="*"){ + positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex)#positionsWithinCodon[(ONEmerCodonPos[1]%%3)+1] + BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + probNonStopMutations[ONEmerReadingFrameCodon,positionWithinCodon]) + } + } + }else{ # Only silent mutations + if( !any(grep("N",ONEmerReadingFrameCodonInputSeq)) & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)!="*" & translateCodonToAminoAcid(ONEmerReadingFrameCodonInputSeq)==translateCodonToAminoAcid(ONEmerReadingFrameCodon) ){ + positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex) + BackgroundMatrix[ONEmer] <- (BackgroundMatrix[ONEmer] + probSMutations[ONEmerReadingFrameCodon,positionWithinCodon]) + } + } + } + } + } + + #Mutations + if(stopMutations==1) mutationCount = mutationCount[mutationCount!="Stop"] + if(model==1) mutationCount = mutationCount[mutationCount=="S"] + mutationPositions = as.numeric(names(mutationCount)) + mutationCount = mutationCount[mutationPositions>2 & mutationPositions<(seqLength-1)] + mutationPositions = mutationPositions[mutationPositions>2 & mutationPositions<(seqLength-1)] + countMutations = 0 + for(mutationPosition in mutationPositions){ + ONEmerIndex = mutationPosition + ONEmer = ONEmerSeq[ONEmerIndex] + GLONEmer = ONEmerGL[ONEmerIndex] + ONEmerCodonPos = getCodonPos(ONEmerIndex) + ONEmerReadingFrameCodon = c2s(ONEmerSeq[ONEmerCodonPos]) + ONEmerReadingFrameCodonGL =c2s(ONEmerGL[ONEmerCodonPos]) + if(!any(grep("N",ONEmer)) & !any(grep("N",GLONEmer))){ + if(model==0){ + countMutations = countMutations + 1 + MutationMatrix[GLONEmer] <- (MutationMatrix[GLONEmer] + 1) + MutationCountMatrix[GLONEmer] <- (MutationCountMatrix[GLONEmer] + 1) + }else{ + if( translateCodonToAminoAcid(ONEmerReadingFrameCodonGL)!="*" ){ + countMutations = countMutations + 1 + positionWithinCodon = which(ONEmerCodonPos==ONEmerIndex) + glNuc = substr(ONEmerReadingFrameCodonGL,positionWithinCodon,positionWithinCodon) + inputNuc = substr(ONEmerReadingFrameCodon,positionWithinCodon,positionWithinCodon) + MutationMatrix[GLONEmer] <- (MutationMatrix[GLONEmer] + substitution[glNuc,inputNuc]) + MutationCountMatrix[GLONEmer] <- (MutationCountMatrix[GLONEmer] + 1) + } + } + } + } + + seqMutability = MutationMatrix/BackgroundMatrix + seqMutability = seqMutability/sum(seqMutability,na.rm=TRUE) + #cat(inputMatrixIndex,"\t",countMutations,"\n") + return(list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix, "BackgroundMatrix"=BackgroundMatrix)) +# tmp<-list("seqMutability" = seqMutability,"numbMutations" = countMutations,"seqMutabilityCount" = MutationCountMatrix) + } + } + +################ +# $Id: trim.R 989 2006-10-29 15:28:26Z ggorjan $ + +trim <- function(s, recode.factor=TRUE, ...) + UseMethod("trim", s) + +trim.default <- function(s, recode.factor=TRUE, ...) + s + +trim.character <- function(s, recode.factor=TRUE, ...) +{ + s <- sub(pattern="^ +", replacement="", x=s) + s <- sub(pattern=" +$", replacement="", x=s) + s +} + +trim.factor <- function(s, recode.factor=TRUE, ...) +{ + levels(s) <- trim(levels(s)) + if(recode.factor) { + dots <- list(x=s, ...) + if(is.null(dots$sort)) dots$sort <- sort + s <- do.call(what=reorder.factor, args=dots) + } + s +} + +trim.list <- function(s, recode.factor=TRUE, ...) + lapply(s, trim, recode.factor=recode.factor, ...) + +trim.data.frame <- function(s, recode.factor=TRUE, ...) +{ + s[] <- trim.list(s, recode.factor=recode.factor, ...) + s +} +####################################### +# Compute the expected for each sequence-germline pair by codon +getExpectedIndividualByCodon <- function(matInput){ +if( any(grep("multicore",search())) ){ + facGL <- factor(matInput[,2]) + facLevels = levels(facGL) + LisGLs_MutabilityU = mclapply(1:length(facLevels), function(x){ + computeMutabilities(facLevels[x]) + }) + facIndex = match(facGL,facLevels) + + LisGLs_Mutability = mclapply(1:nrow(matInput), function(x){ + cInput = rep(NA,nchar(matInput[x,1])) + cInput[s2c(matInput[x,1])!="N"] = 1 + LisGLs_MutabilityU[[facIndex[x]]] * cInput + }) + + LisGLs_Targeting = mclapply(1:dim(matInput)[1], function(x){ + computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) + }) + + LisGLs_MutationTypes = mclapply(1:length(matInput[,2]),function(x){ + #print(x) + computeMutationTypes(matInput[x,2]) + }) + + LisGLs_R_Exp = mclapply(1:nrow(matInput), function(x){ + Exp_R <- rollapply(as.zoo(1:readEnd),width=3,by=3, + function(codonNucs){ + RPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="R") + sum( LisGLs_Targeting[[x]][,codonNucs][RPos], na.rm=T ) + } + ) + }) + + LisGLs_S_Exp = mclapply(1:nrow(matInput), function(x){ + Exp_S <- rollapply(as.zoo(1:readEnd),width=3,by=3, + function(codonNucs){ + SPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="S") + sum( LisGLs_Targeting[[x]][,codonNucs][SPos], na.rm=T ) + } + ) + }) + + Exp_R = matrix(unlist(LisGLs_R_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) + Exp_S = matrix(unlist(LisGLs_S_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) + return( list( "Expected_R"=Exp_R, "Expected_S"=Exp_S) ) + }else{ + facGL <- factor(matInput[,2]) + facLevels = levels(facGL) + LisGLs_MutabilityU = lapply(1:length(facLevels), function(x){ + computeMutabilities(facLevels[x]) + }) + facIndex = match(facGL,facLevels) + + LisGLs_Mutability = lapply(1:nrow(matInput), function(x){ + cInput = rep(NA,nchar(matInput[x,1])) + cInput[s2c(matInput[x,1])!="N"] = 1 + LisGLs_MutabilityU[[facIndex[x]]] * cInput + }) + + LisGLs_Targeting = lapply(1:dim(matInput)[1], function(x){ + computeTargeting(matInput[x,2],LisGLs_Mutability[[x]]) + }) + + LisGLs_MutationTypes = lapply(1:length(matInput[,2]),function(x){ + #print(x) + computeMutationTypes(matInput[x,2]) + }) + + LisGLs_R_Exp = lapply(1:nrow(matInput), function(x){ + Exp_R <- rollapply(as.zoo(1:readEnd),width=3,by=3, + function(codonNucs){ + RPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="R") + sum( LisGLs_Targeting[[x]][,codonNucs][RPos], na.rm=T ) + } + ) + }) + + LisGLs_S_Exp = lapply(1:nrow(matInput), function(x){ + Exp_S <- rollapply(as.zoo(1:readEnd),width=3,by=3, + function(codonNucs){ + SPos = which(LisGLs_MutationTypes[[x]][,codonNucs]=="S") + sum( LisGLs_Targeting[[x]][,codonNucs][SPos], na.rm=T ) + } + ) + }) + + Exp_R = matrix(unlist(LisGLs_R_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) + Exp_S = matrix(unlist(LisGLs_S_Exp),nrow=nrow(matInput),ncol=readEnd/3,T) + return( list( "Expected_R"=Exp_R, "Expected_S"=Exp_S) ) + } +} + +# getObservedMutationsByCodon <- function(listMutations){ +# numbSeqs <- length(listMutations) +# obsMu_R <- matrix(0,nrow=numbSeqs,ncol=readEnd/3,dimnames=list(c(1:numbSeqs),c(1:(readEnd/3)))) +# obsMu_S <- obsMu_R +# temp <- mclapply(1:length(listMutations), function(i){ +# arrMutations = listMutations[[i]] +# RPos = as.numeric(names(arrMutations)[arrMutations=="R"]) +# RPos <- sapply(RPos,getCodonNumb) +# if(any(RPos)){ +# tabR <- table(RPos) +# obsMu_R[i,as.numeric(names(tabR))] <<- tabR +# } +# +# SPos = as.numeric(names(arrMutations)[arrMutations=="S"]) +# SPos <- sapply(SPos,getCodonNumb) +# if(any(SPos)){ +# tabS <- table(SPos) +# obsMu_S[i,names(tabS)] <<- tabS +# } +# } +# ) +# return( list( "Observed_R"=obsMu_R, "Observed_S"=obsMu_S) ) +# } + +getObservedMutationsByCodon <- function(listMutations){ + numbSeqs <- length(listMutations) + obsMu_R <- matrix(0,nrow=numbSeqs,ncol=readEnd/3,dimnames=list(c(1:numbSeqs),c(1:(readEnd/3)))) + obsMu_S <- obsMu_R + temp <- lapply(1:length(listMutations), function(i){ + arrMutations = listMutations[[i]] + RPos = as.numeric(names(arrMutations)[arrMutations=="R"]) + RPos <- sapply(RPos,getCodonNumb) + if(any(RPos)){ + tabR <- table(RPos) + obsMu_R[i,as.numeric(names(tabR))] <<- tabR + } + + SPos = as.numeric(names(arrMutations)[arrMutations=="S"]) + SPos <- sapply(SPos,getCodonNumb) + if(any(SPos)){ + tabS <- table(SPos) + obsMu_S[i,names(tabS)] <<- tabS + } + } + ) + return( list( "Observed_R"=obsMu_R, "Observed_S"=obsMu_S) ) +} + diff -r 43a1aa648537 -r ba33b94637ca baseline/Baseline_Main.r --- a/baseline/Baseline_Main.r Thu Dec 07 03:44:38 2017 -0500 +++ b/baseline/Baseline_Main.r Tue Jan 29 03:54:09 2019 -0500 @@ -1,388 +1,388 @@ -######################################################################################### -# License Agreement -# -# THIS WORK IS PROVIDED UNDER THE TERMS OF THIS CREATIVE COMMONS PUBLIC LICENSE -# ("CCPL" OR "LICENSE"). THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER -# APPLICABLE LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UNDER THIS LICENSE -# OR COPYRIGHT LAW IS PROHIBITED. -# -# BY EXERCISING ANY RIGHTS TO THE WORK PROVIDED HERE, YOU ACCEPT AND AGREE TO BE -# BOUND BY THE TERMS OF THIS LICENSE. TO THE EXTENT THIS LICENSE MAY BE CONSIDERED -# TO BE A CONTRACT, THE LICENSOR GRANTS YOU THE RIGHTS CONTAINED HERE IN -# CONSIDERATION OF YOUR ACCEPTANCE OF SUCH TERMS AND CONDITIONS. -# -# BASELIne: Bayesian Estimation of Antigen-Driven Selection in Immunoglobulin Sequences -# Coded by: Mohamed Uduman & Gur Yaari -# Copyright 2012 Kleinstein Lab -# Version: 1.3 (01/23/2014) -######################################################################################### - -op <- options(); -options(showWarnCalls=FALSE, showErrorCalls=FALSE, warn=-1) -library('seqinr') -if( F & Sys.info()[1]=="Linux"){ - library("multicore") -} - -# Load functions and initialize global variables -source("Baseline_Functions.r") - -# Initialize parameters with user provided arguments - arg <- commandArgs(TRUE) - #arg = c(2,1,5,5,0,1,"1:26:38:55:65:104:116", "test.fasta","","sample") - #arg = c(1,1,5,5,0,1,"1:38:55:65:104:116:200", "test.fasta","","sample") - #arg = c(1,1,5,5,1,1,"1:26:38:55:65:104:116", "/home/mu37/Wu/Wu_Cloned_gapped_sequences_D-masked.fasta","/home/mu37/Wu/","Wu") - testID <- as.numeric(arg[1]) # 1 = Focused, 2 = Local - species <- as.numeric(arg[2]) # 1 = Human. 2 = Mouse - substitutionModel <- as.numeric(arg[3]) # 0 = Uniform substitution, 1 = Smith DS et al. 1996, 5 = FiveS - mutabilityModel <- as.numeric(arg[4]) # 0 = Uniform mutablity, 1 = Tri-nucleotide (Shapiro GS et al. 2002) , 5 = FiveS - clonal <- as.numeric(arg[5]) # 0 = Independent sequences, 1 = Clonally related, 2 = Clonally related & only non-terminal mutations - fixIndels <- as.numeric(arg[6]) # 0 = Do nothing, 1 = Try and fix Indels - region <- as.numeric(strsplit(arg[7],":")[[1]]) # StartPos:LastNucleotideF1:C1:F2:C2:F3:C3 - inputFilePath <- arg[8] # Full path to input file - outputPath <- arg[9] # Full path to location of output files - outputID <- arg[10] # ID for session output - - - if(testID==5){ - traitChangeModel <- 1 - if( !is.na(any(arg[11])) ) traitChangeModel <- as.numeric(arg[11]) # 1 <- Chothia 1998 - initializeTraitChange(traitChangeModel) - } - -# Initialize other parameters/variables - - # Initialzie the codon table ( definitions of R/S ) - computeCodonTable(testID) - - # Initialize - # Test Name - testName<-"Focused" - if(testID==2) testName<-"Local" - if(testID==3) testName<-"Imbalanced" - if(testID==4) testName<-"ImbalancedSilent" - - # Indel placeholders initialization - indelPos <- NULL - delPos <- NULL - insPos <- NULL - - # Initialize in Tranistion & Mutability matrixes - substitution <- initializeSubstitutionMatrix(substitutionModel,species) - mutability <- initializeMutabilityMatrix(mutabilityModel,species) - - # FWR/CDR boundaries - flagTrim <- F - if( is.na(region[7])){ - flagTrim <- T - region[7]<-region[6] - } - readStart = min(region,na.rm=T) - readEnd = max(region,na.rm=T) - if(readStart>1){ - region = region - (readStart - 1) - } - region_Nuc = c( (region[1]*3-2) , (region[2:7]*3) ) - region_Cod = region - - readStart = (readStart*3)-2 - readEnd = (readEnd*3) - - FWR_Nuc <- c( rep(TRUE,(region_Nuc[2])), - rep(FALSE,(region_Nuc[3]-region_Nuc[2])), - rep(TRUE,(region_Nuc[4]-region_Nuc[3])), - rep(FALSE,(region_Nuc[5]-region_Nuc[4])), - rep(TRUE,(region_Nuc[6]-region_Nuc[5])), - rep(FALSE,(region_Nuc[7]-region_Nuc[6])) - ) - CDR_Nuc <- (1-FWR_Nuc) - CDR_Nuc <- as.logical(CDR_Nuc) - FWR_Nuc_Mat <- matrix( rep(FWR_Nuc,4), ncol=length(FWR_Nuc), nrow=4, byrow=T) - CDR_Nuc_Mat <- matrix( rep(CDR_Nuc,4), ncol=length(CDR_Nuc), nrow=4, byrow=T) - - FWR_Codon <- c( rep(TRUE,(region[2])), - rep(FALSE,(region[3]-region[2])), - rep(TRUE,(region[4]-region[3])), - rep(FALSE,(region[5]-region[4])), - rep(TRUE,(region[6]-region[5])), - rep(FALSE,(region[7]-region[6])) - ) - CDR_Codon <- (1-FWR_Codon) - CDR_Codon <- as.logical(CDR_Codon) - - -# Read input FASTA file - tryCatch( - inputFASTA <- baseline.read.fasta(inputFilePath, seqtype="DNA",as.string=T,set.attributes=F,forceDNAtolower=F) - , error = function(ex){ - cat("Error|Error reading input. Please enter or upload a valid FASTA file.\n") - q() - } - ) - - if (length(inputFASTA)==1) { - cat("Error|Error reading input. Please enter or upload a valid FASTA file.\n") - q() - } - - # Process sequence IDs/names - names(inputFASTA) <- sapply(names(inputFASTA),function(x){trim(x)}) - - # Convert non nucleotide characters to N - inputFASTA[length(inputFASTA)] = gsub("\t","",inputFASTA[length(inputFASTA)]) - inputFASTA <- lapply(inputFASTA,replaceNonFASTAChars) - - # Process the FASTA file and conver to Matrix[inputSequence, germlineSequence] - processedInput <- processInputAdvanced(inputFASTA) - matInput <- processedInput[[1]] - germlines <- processedInput[[2]] - lenGermlines = length(unique(germlines)) - groups <- processedInput[[3]] - lenGroups = length(unique(groups)) - rm(processedInput) - rm(inputFASTA) - -# # remove clones with less than 2 seqeunces -# tableGL <- table(germlines) -# singletons <- which(tableGL<8) -# rowsToRemove <- match(singletons,germlines) -# if(any(rowsToRemove)){ -# matInput <- matInput[-rowsToRemove,] -# germlines <- germlines[-rowsToRemove] -# groups <- groups[-rowsToRemove] -# } -# -# # remove unproductive seqs -# nonFuctionalSeqs <- sapply(rownames(matInput),function(x){any(grep("unproductive",x))}) -# if(any(nonFuctionalSeqs)){ -# if(sum(nonFuctionalSeqs)==length(germlines)){ -# write.table("Unproductive",file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) -# q() -# } -# matInput <- matInput[-which(nonFuctionalSeqs),] -# germlines <- germlines[-which(nonFuctionalSeqs)] -# germlines[1:length(germlines)] <- 1:length(germlines) -# groups <- groups[-which(nonFuctionalSeqs)] -# } -# -# if(class(matInput)=="character"){ -# write.table("All unproductive seqs",file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) -# q() -# } -# -# if(nrow(matInput)<10 | is.null(nrow(matInput))){ -# write.table(paste(nrow(matInput), "seqs only",sep=""),file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) -# q() -# } - -# replace leading & trailing "-" with "N: - matInput <- t(apply(matInput,1,replaceLeadingTrailingDashes,readEnd)) - - # Trim (nucleotide) input sequences to the last codon - #matInput[,1] <- apply(matrix(matInput[,1]),1,trimToLastCodon) - -# # Check for Indels -# if(fixIndels){ -# delPos <- fixDeletions(matInput) -# insPos <- fixInsertions(matInput) -# }else{ -# # Check for indels -# indelPos <- checkForInDels(matInput) -# indelPos <- apply(cbind(indelPos[[1]],indelPos[[2]]),1,function(x){(x[1]==T & x[2]==T)}) -# } - - # If indels are present, remove mutations in the seqeunce & throw warning at end - #matInput[indelPos,] <- apply(matrix(matInput[indelPos,],nrow=sum(indelPos),ncol=2),1,function(x){x[1]=x[2]; return(x) }) - - colnames(matInput)=c("Input","Germline") - - # If seqeunces are clonal, create effective sequence for each clone & modify germline/group definitions - germlinesOriginal = NULL - if(clonal){ - germlinesOriginal <- germlines - collapseCloneResults <- tapply(1:nrow(matInput),germlines,function(i){ - collapseClone(matInput[i,1],matInput[i[1],2],readEnd,nonTerminalOnly=(clonal-1)) - }) - matInput = t(sapply(collapseCloneResults,function(x){return(x[[1]])})) - names_groups = tapply(groups,germlines,function(x){names(x[1])}) - groups = tapply(groups,germlines,function(x){array(x[1],dimnames=names(x[1]))}) - names(groups) = names_groups - - names_germlines = tapply(germlines,germlines,function(x){names(x[1])}) - germlines = tapply( germlines,germlines,function(x){array(x[1],dimnames=names(x[1]))} ) - names(germlines) = names_germlines - matInputErrors = sapply(collapseCloneResults,function(x){return(x[[2]])}) - } - - -# Selection Analysis - - -# if (length(germlines)>sequenceLimit) { -# # Code to parallelize processing goes here -# stop( paste("Error: Cannot process more than ", Upper_limit," sequences",sep="") ) -# } - -# if (length(germlines)1){ - groups <- c(groups,lenGroups+1) - names(groups)[length(groups)] = "All sequences combined" - bayesPDF_groups_cdr[[lenGroups+1]] = groupPosteriors(bayesPDF_groups_cdr,length_sigma=4001) - bayesPDF_groups_fwr[[lenGroups+1]] = groupPosteriors(bayesPDF_groups_fwr,length_sigma=4001) - } - - #Bayesian Outputs - bayes_cdr = t(sapply(bayesPDF_cdr,calcBayesOutputInfo)) - bayes_fwr = t(sapply(bayesPDF_fwr,calcBayesOutputInfo)) - bayes_germlines_cdr = t(sapply(bayesPDF_germlines_cdr,calcBayesOutputInfo)) - bayes_germlines_fwr = t(sapply(bayesPDF_germlines_fwr,calcBayesOutputInfo)) - bayes_groups_cdr = t(sapply(bayesPDF_groups_cdr,calcBayesOutputInfo)) - bayes_groups_fwr = t(sapply(bayesPDF_groups_fwr,calcBayesOutputInfo)) - - #P-values - simgaP_cdr = sapply(bayesPDF_cdr,computeSigmaP) - simgaP_fwr = sapply(bayesPDF_fwr,computeSigmaP) - - simgaP_germlines_cdr = sapply(bayesPDF_germlines_cdr,computeSigmaP) - simgaP_germlines_fwr = sapply(bayesPDF_germlines_fwr,computeSigmaP) - - simgaP_groups_cdr = sapply(bayesPDF_groups_cdr,computeSigmaP) - simgaP_groups_fwr = sapply(bayesPDF_groups_fwr,computeSigmaP) - - - #Format output - - # Round expected mutation frequencies to 3 decimal places - matMutationInfo[germlinesOriginal[indelPos],] = NA - if(nrow(matMutationInfo)==1){ - matMutationInfo[5:8] = round(matMutationInfo[,5:8]/sum(matMutationInfo[,5:8],na.rm=T),3) - }else{ - matMutationInfo[,5:8] = t(round(apply(matMutationInfo[,5:8],1,function(x){ return(x/sum(x,na.rm=T)) }),3)) - } - - listPDFs = list() - nRows = length(unique(groups)) + length(unique(germlines)) + length(groups) - - matOutput = matrix(NA,ncol=18,nrow=nRows) - rowNumb = 1 - for(G in unique(groups)){ - #print(G) - matOutput[rowNumb,c(1,2,11:18)] = c("Group",names(groups)[groups==G][1],bayes_groups_cdr[G,],bayes_groups_fwr[G,],simgaP_groups_cdr[G],simgaP_groups_fwr[G]) - listPDFs[[rowNumb]] = list("CDR"=bayesPDF_groups_cdr[[G]],"FWR"=bayesPDF_groups_fwr[[G]]) - names(listPDFs)[rowNumb] = names(groups[groups==paste(G)])[1] - #if(names(groups)[which(groups==G)[1]]!="All sequences combined"){ - gs = unique(germlines[groups==G]) - rowNumb = rowNumb+1 - if( !is.na(gs) ){ - for( g in gs ){ - matOutput[rowNumb,c(1,2,11:18)] = c("Germline",names(germlines)[germlines==g][1],bayes_germlines_cdr[g,],bayes_germlines_fwr[g,],simgaP_germlines_cdr[g],simgaP_germlines_fwr[g]) - listPDFs[[rowNumb]] = list("CDR"=bayesPDF_germlines_cdr[[g]],"FWR"=bayesPDF_germlines_fwr[[g]]) - names(listPDFs)[rowNumb] = names(germlines[germlines==paste(g)])[1] - rowNumb = rowNumb+1 - indexesOfInterest = which(germlines==g) - numbSeqsOfInterest = length(indexesOfInterest) - rowNumb = seq(rowNumb,rowNumb+(numbSeqsOfInterest-1)) - matOutput[rowNumb,] = matrix( c( rep("Sequence",numbSeqsOfInterest), - rownames(matInput)[indexesOfInterest], - c(matMutationInfo[indexesOfInterest,1:4]), - c(matMutationInfo[indexesOfInterest,5:8]), - c(bayes_cdr[indexesOfInterest,]), - c(bayes_fwr[indexesOfInterest,]), - c(simgaP_cdr[indexesOfInterest]), - c(simgaP_fwr[indexesOfInterest]) - ), ncol=18, nrow=numbSeqsOfInterest,byrow=F) - increment=0 - for( ioi in indexesOfInterest){ - listPDFs[[min(rowNumb)+increment]] = list("CDR"=bayesPDF_cdr[[ioi]] , "FWR"=bayesPDF_fwr[[ioi]]) - names(listPDFs)[min(rowNumb)+increment] = rownames(matInput)[ioi] - increment = increment + 1 - } - rowNumb=max(rowNumb)+1 - - } - } - } - colsToFormat = 11:18 - matOutput[,colsToFormat] = formatC( matrix(as.numeric(matOutput[,colsToFormat]), nrow=nrow(matOutput), ncol=length(colsToFormat)) , digits=3) - matOutput[matOutput== " NaN"] = NA - - - - colnames(matOutput) = c("Type", "ID", "Observed_CDR_R", "Observed_CDR_S", "Observed_FWR_R", "Observed_FWR_S", - "Expected_CDR_R", "Expected_CDR_S", "Expected_FWR_R", "Expected_FWR_S", - paste( rep(testName,6), rep(c("Sigma","CIlower","CIupper"),2),rep(c("CDR","FWR"),each=3), sep="_"), - paste( rep(testName,2), rep("P",2),c("CDR","FWR"), sep="_") - ) - fileName = paste(outputPath,outputID,".txt",sep="") - write.table(matOutput,file=fileName,quote=F,sep="\t",row.names=T,col.names=NA) - fileName = paste(outputPath,outputID,".RData",sep="") - save(listPDFs,file=fileName) - -indelWarning = FALSE -if(sum(indelPos)>0){ - indelWarning = "

Warning: The following sequences have either gaps and/or deletions, and have been ommited from the analysis."; - indelWarning = paste( indelWarning , "

    ", sep="" ) - for(indels in names(indelPos)[indelPos]){ - indelWarning = paste( indelWarning , "
  • ", indels, "
    • ", sep="" ) - } - indelWarning = paste( indelWarning , "

", sep="" ) -} - -cloneWarning = FALSE -if(clonal==1){ - if(sum(matInputErrors)>0){ - cloneWarning = "

Warning: The following clones have sequences of unequal length."; - cloneWarning = paste( cloneWarning , "

    ", sep="" ) - for(clone in names(matInputErrors)[matInputErrors]){ - cloneWarning = paste( cloneWarning , "
  • ", names(germlines)[as.numeric(clone)], "
    • ", sep="" ) - } - cloneWarning = paste( cloneWarning , "

", sep="" ) - } -} -cat(paste("Success",outputID,indelWarning,cloneWarning,sep="|")) +######################################################################################### +# License Agreement +# +# THIS WORK IS PROVIDED UNDER THE TERMS OF THIS CREATIVE COMMONS PUBLIC LICENSE +# ("CCPL" OR "LICENSE"). THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER +# APPLICABLE LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UNDER THIS LICENSE +# OR COPYRIGHT LAW IS PROHIBITED. +# +# BY EXERCISING ANY RIGHTS TO THE WORK PROVIDED HERE, YOU ACCEPT AND AGREE TO BE +# BOUND BY THE TERMS OF THIS LICENSE. TO THE EXTENT THIS LICENSE MAY BE CONSIDERED +# TO BE A CONTRACT, THE LICENSOR GRANTS YOU THE RIGHTS CONTAINED HERE IN +# CONSIDERATION OF YOUR ACCEPTANCE OF SUCH TERMS AND CONDITIONS. +# +# BASELIne: Bayesian Estimation of Antigen-Driven Selection in Immunoglobulin Sequences +# Coded by: Mohamed Uduman & Gur Yaari +# Copyright 2012 Kleinstein Lab +# Version: 1.3 (01/23/2014) +######################################################################################### + +op <- options(); +options(showWarnCalls=FALSE, showErrorCalls=FALSE, warn=-1) +library('seqinr') +if( F & Sys.info()[1]=="Linux"){ + library("multicore") +} + +# Load functions and initialize global variables +source("Baseline_Functions.r") + +# Initialize parameters with user provided arguments + arg <- commandArgs(TRUE) + #arg = c(2,1,5,5,0,1,"1:26:38:55:65:104:116", "test.fasta","","sample") + #arg = c(1,1,5,5,0,1,"1:38:55:65:104:116:200", "test.fasta","","sample") + #arg = c(1,1,5,5,1,1,"1:26:38:55:65:104:116", "/home/mu37/Wu/Wu_Cloned_gapped_sequences_D-masked.fasta","/home/mu37/Wu/","Wu") + testID <- as.numeric(arg[1]) # 1 = Focused, 2 = Local + species <- as.numeric(arg[2]) # 1 = Human. 2 = Mouse + substitutionModel <- as.numeric(arg[3]) # 0 = Uniform substitution, 1 = Smith DS et al. 1996, 5 = FiveS + mutabilityModel <- as.numeric(arg[4]) # 0 = Uniform mutablity, 1 = Tri-nucleotide (Shapiro GS et al. 2002) , 5 = FiveS + clonal <- as.numeric(arg[5]) # 0 = Independent sequences, 1 = Clonally related, 2 = Clonally related & only non-terminal mutations + fixIndels <- as.numeric(arg[6]) # 0 = Do nothing, 1 = Try and fix Indels + region <- as.numeric(strsplit(arg[7],":")[[1]]) # StartPos:LastNucleotideF1:C1:F2:C2:F3:C3 + inputFilePath <- arg[8] # Full path to input file + outputPath <- arg[9] # Full path to location of output files + outputID <- arg[10] # ID for session output + + + if(testID==5){ + traitChangeModel <- 1 + if( !is.na(any(arg[11])) ) traitChangeModel <- as.numeric(arg[11]) # 1 <- Chothia 1998 + initializeTraitChange(traitChangeModel) + } + +# Initialize other parameters/variables + + # Initialzie the codon table ( definitions of R/S ) + computeCodonTable(testID) + + # Initialize + # Test Name + testName<-"Focused" + if(testID==2) testName<-"Local" + if(testID==3) testName<-"Imbalanced" + if(testID==4) testName<-"ImbalancedSilent" + + # Indel placeholders initialization + indelPos <- NULL + delPos <- NULL + insPos <- NULL + + # Initialize in Tranistion & Mutability matrixes + substitution <- initializeSubstitutionMatrix(substitutionModel,species) + mutability <- initializeMutabilityMatrix(mutabilityModel,species) + + # FWR/CDR boundaries + flagTrim <- F + if( is.na(region[7])){ + flagTrim <- T + region[7]<-region[6] + } + readStart = min(region,na.rm=T) + readEnd = max(region,na.rm=T) + if(readStart>1){ + region = region - (readStart - 1) + } + region_Nuc = c( (region[1]*3-2) , (region[2:7]*3) ) + region_Cod = region + + readStart = (readStart*3)-2 + readEnd = (readEnd*3) + + FWR_Nuc <- c( rep(TRUE,(region_Nuc[2])), + rep(FALSE,(region_Nuc[3]-region_Nuc[2])), + rep(TRUE,(region_Nuc[4]-region_Nuc[3])), + rep(FALSE,(region_Nuc[5]-region_Nuc[4])), + rep(TRUE,(region_Nuc[6]-region_Nuc[5])), + rep(FALSE,(region_Nuc[7]-region_Nuc[6])) + ) + CDR_Nuc <- (1-FWR_Nuc) + CDR_Nuc <- as.logical(CDR_Nuc) + FWR_Nuc_Mat <- matrix( rep(FWR_Nuc,4), ncol=length(FWR_Nuc), nrow=4, byrow=T) + CDR_Nuc_Mat <- matrix( rep(CDR_Nuc,4), ncol=length(CDR_Nuc), nrow=4, byrow=T) + + FWR_Codon <- c( rep(TRUE,(region[2])), + rep(FALSE,(region[3]-region[2])), + rep(TRUE,(region[4]-region[3])), + rep(FALSE,(region[5]-region[4])), + rep(TRUE,(region[6]-region[5])), + rep(FALSE,(region[7]-region[6])) + ) + CDR_Codon <- (1-FWR_Codon) + CDR_Codon <- as.logical(CDR_Codon) + + +# Read input FASTA file + tryCatch( + inputFASTA <- baseline.read.fasta(inputFilePath, seqtype="DNA",as.string=T,set.attributes=F,forceDNAtolower=F) + , error = function(ex){ + cat("Error|Error reading input. Please enter or upload a valid FASTA file.\n") + q() + } + ) + + if (length(inputFASTA)==1) { + cat("Error|Error reading input. Please enter or upload a valid FASTA file.\n") + q() + } + + # Process sequence IDs/names + names(inputFASTA) <- sapply(names(inputFASTA),function(x){trim(x)}) + + # Convert non nucleotide characters to N + inputFASTA[length(inputFASTA)] = gsub("\t","",inputFASTA[length(inputFASTA)]) + inputFASTA <- lapply(inputFASTA,replaceNonFASTAChars) + + # Process the FASTA file and conver to Matrix[inputSequence, germlineSequence] + processedInput <- processInputAdvanced(inputFASTA) + matInput <- processedInput[[1]] + germlines <- processedInput[[2]] + lenGermlines = length(unique(germlines)) + groups <- processedInput[[3]] + lenGroups = length(unique(groups)) + rm(processedInput) + rm(inputFASTA) + +# # remove clones with less than 2 seqeunces +# tableGL <- table(germlines) +# singletons <- which(tableGL<8) +# rowsToRemove <- match(singletons,germlines) +# if(any(rowsToRemove)){ +# matInput <- matInput[-rowsToRemove,] +# germlines <- germlines[-rowsToRemove] +# groups <- groups[-rowsToRemove] +# } +# +# # remove unproductive seqs +# nonFuctionalSeqs <- sapply(rownames(matInput),function(x){any(grep("unproductive",x))}) +# if(any(nonFuctionalSeqs)){ +# if(sum(nonFuctionalSeqs)==length(germlines)){ +# write.table("Unproductive",file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) +# q() +# } +# matInput <- matInput[-which(nonFuctionalSeqs),] +# germlines <- germlines[-which(nonFuctionalSeqs)] +# germlines[1:length(germlines)] <- 1:length(germlines) +# groups <- groups[-which(nonFuctionalSeqs)] +# } +# +# if(class(matInput)=="character"){ +# write.table("All unproductive seqs",file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) +# q() +# } +# +# if(nrow(matInput)<10 | is.null(nrow(matInput))){ +# write.table(paste(nrow(matInput), "seqs only",sep=""),file=paste(outputPath,outputID,".txt",sep=""),quote=F,sep="\t",row.names=F,col.names=T) +# q() +# } + +# replace leading & trailing "-" with "N: + matInput <- t(apply(matInput,1,replaceLeadingTrailingDashes,readEnd)) + + # Trim (nucleotide) input sequences to the last codon + #matInput[,1] <- apply(matrix(matInput[,1]),1,trimToLastCodon) + +# # Check for Indels +# if(fixIndels){ +# delPos <- fixDeletions(matInput) +# insPos <- fixInsertions(matInput) +# }else{ +# # Check for indels +# indelPos <- checkForInDels(matInput) +# indelPos <- apply(cbind(indelPos[[1]],indelPos[[2]]),1,function(x){(x[1]==T & x[2]==T)}) +# } + + # If indels are present, remove mutations in the seqeunce & throw warning at end + #matInput[indelPos,] <- apply(matrix(matInput[indelPos,],nrow=sum(indelPos),ncol=2),1,function(x){x[1]=x[2]; return(x) }) + + colnames(matInput)=c("Input","Germline") + + # If seqeunces are clonal, create effective sequence for each clone & modify germline/group definitions + germlinesOriginal = NULL + if(clonal){ + germlinesOriginal <- germlines + collapseCloneResults <- tapply(1:nrow(matInput),germlines,function(i){ + collapseClone(matInput[i,1],matInput[i[1],2],readEnd,nonTerminalOnly=(clonal-1)) + }) + matInput = t(sapply(collapseCloneResults,function(x){return(x[[1]])})) + names_groups = tapply(groups,germlines,function(x){names(x[1])}) + groups = tapply(groups,germlines,function(x){array(x[1],dimnames=names(x[1]))}) + names(groups) = names_groups + + names_germlines = tapply(germlines,germlines,function(x){names(x[1])}) + germlines = tapply( germlines,germlines,function(x){array(x[1],dimnames=names(x[1]))} ) + names(germlines) = names_germlines + matInputErrors = sapply(collapseCloneResults,function(x){return(x[[2]])}) + } + + +# Selection Analysis + + +# if (length(germlines)>sequenceLimit) { +# # Code to parallelize processing goes here +# stop( paste("Error: Cannot process more than ", Upper_limit," sequences",sep="") ) +# } + +# if (length(germlines)1){ + groups <- c(groups,lenGroups+1) + names(groups)[length(groups)] = "All sequences combined" + bayesPDF_groups_cdr[[lenGroups+1]] = groupPosteriors(bayesPDF_groups_cdr,length_sigma=4001) + bayesPDF_groups_fwr[[lenGroups+1]] = groupPosteriors(bayesPDF_groups_fwr,length_sigma=4001) + } + + #Bayesian Outputs + bayes_cdr = t(sapply(bayesPDF_cdr,calcBayesOutputInfo)) + bayes_fwr = t(sapply(bayesPDF_fwr,calcBayesOutputInfo)) + bayes_germlines_cdr = t(sapply(bayesPDF_germlines_cdr,calcBayesOutputInfo)) + bayes_germlines_fwr = t(sapply(bayesPDF_germlines_fwr,calcBayesOutputInfo)) + bayes_groups_cdr = t(sapply(bayesPDF_groups_cdr,calcBayesOutputInfo)) + bayes_groups_fwr = t(sapply(bayesPDF_groups_fwr,calcBayesOutputInfo)) + + #P-values + simgaP_cdr = sapply(bayesPDF_cdr,computeSigmaP) + simgaP_fwr = sapply(bayesPDF_fwr,computeSigmaP) + + simgaP_germlines_cdr = sapply(bayesPDF_germlines_cdr,computeSigmaP) + simgaP_germlines_fwr = sapply(bayesPDF_germlines_fwr,computeSigmaP) + + simgaP_groups_cdr = sapply(bayesPDF_groups_cdr,computeSigmaP) + simgaP_groups_fwr = sapply(bayesPDF_groups_fwr,computeSigmaP) + + + #Format output + + # Round expected mutation frequencies to 3 decimal places + matMutationInfo[germlinesOriginal[indelPos],] = NA + if(nrow(matMutationInfo)==1){ + matMutationInfo[5:8] = round(matMutationInfo[,5:8]/sum(matMutationInfo[,5:8],na.rm=T),3) + }else{ + matMutationInfo[,5:8] = t(round(apply(matMutationInfo[,5:8],1,function(x){ return(x/sum(x,na.rm=T)) }),3)) + } + + listPDFs = list() + nRows = length(unique(groups)) + length(unique(germlines)) + length(groups) + + matOutput = matrix(NA,ncol=18,nrow=nRows) + rowNumb = 1 + for(G in unique(groups)){ + #print(G) + matOutput[rowNumb,c(1,2,11:18)] = c("Group",names(groups)[groups==G][1],bayes_groups_cdr[G,],bayes_groups_fwr[G,],simgaP_groups_cdr[G],simgaP_groups_fwr[G]) + listPDFs[[rowNumb]] = list("CDR"=bayesPDF_groups_cdr[[G]],"FWR"=bayesPDF_groups_fwr[[G]]) + names(listPDFs)[rowNumb] = names(groups[groups==paste(G)])[1] + #if(names(groups)[which(groups==G)[1]]!="All sequences combined"){ + gs = unique(germlines[groups==G]) + rowNumb = rowNumb+1 + if( !is.na(gs) ){ + for( g in gs ){ + matOutput[rowNumb,c(1,2,11:18)] = c("Germline",names(germlines)[germlines==g][1],bayes_germlines_cdr[g,],bayes_germlines_fwr[g,],simgaP_germlines_cdr[g],simgaP_germlines_fwr[g]) + listPDFs[[rowNumb]] = list("CDR"=bayesPDF_germlines_cdr[[g]],"FWR"=bayesPDF_germlines_fwr[[g]]) + names(listPDFs)[rowNumb] = names(germlines[germlines==paste(g)])[1] + rowNumb = rowNumb+1 + indexesOfInterest = which(germlines==g) + numbSeqsOfInterest = length(indexesOfInterest) + rowNumb = seq(rowNumb,rowNumb+(numbSeqsOfInterest-1)) + matOutput[rowNumb,] = matrix( c( rep("Sequence",numbSeqsOfInterest), + rownames(matInput)[indexesOfInterest], + c(matMutationInfo[indexesOfInterest,1:4]), + c(matMutationInfo[indexesOfInterest,5:8]), + c(bayes_cdr[indexesOfInterest,]), + c(bayes_fwr[indexesOfInterest,]), + c(simgaP_cdr[indexesOfInterest]), + c(simgaP_fwr[indexesOfInterest]) + ), ncol=18, nrow=numbSeqsOfInterest,byrow=F) + increment=0 + for( ioi in indexesOfInterest){ + listPDFs[[min(rowNumb)+increment]] = list("CDR"=bayesPDF_cdr[[ioi]] , "FWR"=bayesPDF_fwr[[ioi]]) + names(listPDFs)[min(rowNumb)+increment] = rownames(matInput)[ioi] + increment = increment + 1 + } + rowNumb=max(rowNumb)+1 + + } + } + } + colsToFormat = 11:18 + matOutput[,colsToFormat] = formatC( matrix(as.numeric(matOutput[,colsToFormat]), nrow=nrow(matOutput), ncol=length(colsToFormat)) , digits=3) + matOutput[matOutput== " NaN"] = NA + + + + colnames(matOutput) = c("Type", "ID", "Observed_CDR_R", "Observed_CDR_S", "Observed_FWR_R", "Observed_FWR_S", + "Expected_CDR_R", "Expected_CDR_S", "Expected_FWR_R", "Expected_FWR_S", + paste( rep(testName,6), rep(c("Sigma","CIlower","CIupper"),2),rep(c("CDR","FWR"),each=3), sep="_"), + paste( rep(testName,2), rep("P",2),c("CDR","FWR"), sep="_") + ) + fileName = paste(outputPath,outputID,".txt",sep="") + write.table(matOutput,file=fileName,quote=F,sep="\t",row.names=T,col.names=NA) + fileName = paste(outputPath,outputID,".RData",sep="") + save(listPDFs,file=fileName) + +indelWarning = FALSE +if(sum(indelPos)>0){ + indelWarning = "

Warning: The following sequences have either gaps and/or deletions, and have been ommited from the analysis."; + indelWarning = paste( indelWarning , "

    ", sep="" ) + for(indels in names(indelPos)[indelPos]){ + indelWarning = paste( indelWarning , "
  • ", indels, "
    • ", sep="" ) + } + indelWarning = paste( indelWarning , "

", sep="" ) +} + +cloneWarning = FALSE +if(clonal==1){ + if(sum(matInputErrors)>0){ + cloneWarning = "

Warning: The following clones have sequences of unequal length."; + cloneWarning = paste( cloneWarning , "

    ", sep="" ) + for(clone in names(matInputErrors)[matInputErrors]){ + cloneWarning = paste( cloneWarning , "
  • ", names(germlines)[as.numeric(clone)], "
    • ", sep="" ) + } + cloneWarning = paste( cloneWarning , "

", sep="" ) + } +} +cat(paste("Success",outputID,indelWarning,cloneWarning,sep="|")) diff -r 43a1aa648537 -r ba33b94637ca baseline/FiveS_Mutability.RData diff -r 43a1aa648537 -r ba33b94637ca baseline/FiveS_Substitution.RData diff -r 43a1aa648537 -r ba33b94637ca baseline/IMGT-reference-seqs-IGHV-2015-11-05.fa diff -r 43a1aa648537 -r ba33b94637ca baseline/IMGTVHreferencedataset20161215.fa diff -r 43a1aa648537 -r ba33b94637ca baseline/baseline_url.txt --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/baseline/baseline_url.txt Tue Jan 29 03:54:09 2019 -0500 @@ -0,0 +1,1 @@ +http://selection.med.yale.edu/baseline/ \ No newline at end of file diff -r 43a1aa648537 -r ba33b94637ca baseline/comparePDFs.r --- a/baseline/comparePDFs.r Thu Dec 07 03:44:38 2017 -0500 +++ b/baseline/comparePDFs.r Tue Jan 29 03:54:09 2019 -0500 @@ -1,225 +1,225 @@ -options("warn"=-1) - -#from http://selection.med.yale.edu/baseline/Archive/Baseline%20Version%201.3/Baseline_Functions_Version1.3.r -# Compute p-value of two distributions -compareTwoDistsFaster <-function(sigma_S=seq(-20,20,length.out=4001), N=10000, dens1=runif(4001,0,1), dens2=runif(4001,0,1)){ -#print(c(length(dens1),length(dens2))) -if(length(dens1)>1 & length(dens2)>1 ){ - dens1<-dens1/sum(dens1) - dens2<-dens2/sum(dens2) - cum2 <- cumsum(dens2)-dens2/2 - tmp<- sum(sapply(1:length(dens1),function(i)return(dens1[i]*cum2[i]))) - #print(tmp) - if(tmp>0.5)tmp<-tmp-1 - return( tmp ) - } - else { - return(NA) - } - #return (sum(sapply(1:N,function(i)(sample(sigma_S,1,prob=dens1)>sample(sigma_S,1,prob=dens2))))/N) -} - - -require("grid") -arg <- commandArgs(TRUE) -#arg <- c("300143","4","5") -arg[!arg=="clonal"] -input <- arg[1] -output <- arg[2] -rowIDs <- as.numeric( sapply(arg[3:(max(3,length(arg)))],function(x){ gsub("chkbx","",x) } ) ) - -numbSeqs = length(rowIDs) - -if ( is.na(rowIDs[1]) | numbSeqs>10 ) { - stop( paste("Error: Please select between one and 10 seqeunces to compare.") ) -} - -#load( paste("output/",sessionID,".RData",sep="") ) -load( input ) -#input - -xMarks = seq(-20,20,length.out=4001) - -plot_grid_s<-function(pdf1,pdf2,Sample=100,cex=1,xlim=NULL,xMarks = seq(-20,20,length.out=4001)){ - yMax = max(c(abs(as.numeric(unlist(listPDFs[pdf1]))),abs(as.numeric(unlist(listPDFs[pdf2]))),0),na.rm=T) * 1.1 - - if(length(xlim==2)){ - xMin=xlim[1] - xMax=xlim[2] - } else { - xMin_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][1] - xMin_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][1] - xMax_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001])] - xMax_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001])] - - xMin_CDR2 = xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001][1] - xMin_FWR2 = xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001][1] - xMax_CDR2 = xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001])] - xMax_FWR2 = xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001])] - - xMin=min(c(xMin_CDR,xMin_FWR,xMin_CDR2,xMin_FWR2,0),na.rm=TRUE) - xMax=max(c(xMax_CDR,xMax_FWR,xMax_CDR2,xMax_FWR2,0),na.rm=TRUE) - } - - sigma<-approx(xMarks,xout=seq(xMin,xMax,length.out=Sample))$x - grid.rect(gp = gpar(col=gray(0.6),fill="white",cex=cex)) - x <- sigma - pushViewport(viewport(x=0.175,y=0.175,width=0.825,height=0.825,just=c("left","bottom"),default.units="npc")) - #pushViewport(plotViewport(c(1.8, 1.8, 0.25, 0.25)*cex)) - pushViewport(dataViewport(x, c(yMax,-yMax),gp = gpar(cex=cex),extension=c(0.05))) - grid.polygon(c(0,0,1,1),c(0,0.5,0.5,0),gp=gpar(col=grey(0.95),fill=grey(0.95)),default.units="npc") - grid.polygon(c(0,0,1,1),c(1,0.5,0.5,1),gp=gpar(col=grey(0.9),fill=grey(0.9)),default.units="npc") - grid.rect() - grid.xaxis(gp = gpar(cex=cex/1.1)) - yticks = pretty(c(-yMax,yMax),8) - yticks = yticks[yticks>(-yMax) & yticks<(yMax)] - grid.yaxis(at=yticks,label=abs(yticks),gp = gpar(cex=cex/1.1)) - if(length(listPDFs[pdf1][[1]][["CDR"]])>1){ - ycdr<-approx(xMarks,listPDFs[pdf1][[1]][["CDR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y - grid.lines(unit(x,"native"), unit(ycdr,"native"),gp=gpar(col=2,lwd=2)) - } - if(length(listPDFs[pdf1][[1]][["FWR"]])>1){ - yfwr<-approx(xMarks,listPDFs[pdf1][[1]][["FWR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y - grid.lines(unit(x,"native"), unit(-yfwr,"native"),gp=gpar(col=4,lwd=2)) - } - - if(length(listPDFs[pdf2][[1]][["CDR"]])>1){ - ycdr2<-approx(xMarks,listPDFs[pdf2][[1]][["CDR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y - grid.lines(unit(x,"native"), unit(ycdr2,"native"),gp=gpar(col=2,lwd=2,lty=2)) - } - if(length(listPDFs[pdf2][[1]][["FWR"]])>1){ - yfwr2<-approx(xMarks,listPDFs[pdf2][[1]][["FWR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y - grid.lines(unit(x,"native"), unit(-yfwr2,"native"),gp=gpar(col=4,lwd=2,lty=2)) - } - - grid.lines(unit(c(0,1),"npc"), unit(c(0.5,0.5),"npc"),gp=gpar(col=1)) - grid.lines(unit(c(0,0),"native"), unit(c(0,1),"npc"),gp=gpar(col=1,lwd=1,lty=3)) - - grid.text("All", x = unit(-2.5, "lines"), rot = 90,gp = gpar(cex=cex)) - grid.text( expression(paste("Selection Strength (", Sigma, ")", sep="")) , y = unit(-2.5, "lines"),gp = gpar(cex=cex)) - - if(pdf1==pdf2 & length(listPDFs[pdf2][[1]][["FWR"]])>1 & length(listPDFs[pdf2][[1]][["CDR"]])>1 ){ - pCDRFWR = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["CDR"]], dens2=listPDFs[[pdf1]][["FWR"]]) - pval = formatC(as.numeric(pCDRFWR),digits=3) - grid.text( substitute(expression(paste(P[CDR/FWR], "=", x, sep="")),list(x=pval))[[2]] , x = unit(0.02, "npc"),y = unit(0.98, "npc"),just=c("left", "top"),gp = gpar(cex=cex*1.2)) - } - grid.text(paste("CDR"), x = unit(0.98, "npc"),y = unit(0.98, "npc"),just=c("right", "top"),gp = gpar(cex=cex*1.5)) - grid.text(paste("FWR"), x = unit(0.98, "npc"),y = unit(0.02, "npc"),just=c("right", "bottom"),gp = gpar(cex=cex*1.5)) - popViewport(2) -} -#plot_grid_s(1) - - -p2col<-function(p=0.01){ - breaks=c(-.51,-0.1,-.05,-0.01,-0.005,0,0.005,0.01,0.05,0.1,0.51) - i<-findInterval(p,breaks) - cols = c( rgb(0.8,1,0.8), rgb(0.6,1,0.6), rgb(0.4,1,0.4), rgb(0.2,1,0.2) , rgb(0,1,0), - rgb(1,0,0), rgb(1,.2,.2), rgb(1,.4,.4), rgb(1,.6,.6) , rgb(1,.8,.8) ) - return(cols[i]) -} - - -plot_pvals<-function(pdf1,pdf2,cex=1,upper=TRUE){ - if(upper){ - pCDR1FWR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["CDR"]], dens2=listPDFs[[pdf2]][["FWR"]]) - pFWR1FWR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["FWR"]], dens2=listPDFs[[pdf2]][["FWR"]]) - pFWR1CDR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens2=listPDFs[[pdf2]][["CDR"]], dens1=listPDFs[[pdf1]][["FWR"]]) - pCDR1CDR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens2=listPDFs[[pdf2]][["CDR"]], dens1=listPDFs[[pdf1]][["CDR"]]) - grid.polygon(c(0.5,0.5,1,1),c(0,0.5,0.5,0),gp=gpar(col=p2col(pFWR1FWR2),fill=p2col(pFWR1FWR2)),default.units="npc") - grid.polygon(c(0.5,0.5,1,1),c(1,0.5,0.5,1),gp=gpar(col=p2col(pCDR1FWR2),fill=p2col(pCDR1FWR2)),default.units="npc") - grid.polygon(c(0.5,0.5,0,0),c(1,0.5,0.5,1),gp=gpar(col=p2col(pCDR1CDR2),fill=p2col(pCDR1CDR2)),default.units="npc") - grid.polygon(c(0.5,0.5,0,0),c(0,0.5,0.5,0),gp=gpar(col=p2col(pFWR1CDR2),fill=p2col(pFWR1CDR2)),default.units="npc") - - grid.lines(c(0,1),0.5,gp=gpar(lty=2,col=gray(0.925))) - grid.lines(0.5,c(0,1),gp=gpar(lty=2,col=gray(0.925))) - - grid.text(formatC(as.numeric(pFWR1FWR2),digits=3), x = unit(0.75, "npc"),y = unit(0.25, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) - grid.text(formatC(as.numeric(pCDR1FWR2),digits=3), x = unit(0.75, "npc"),y = unit(0.75, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) - grid.text(formatC(as.numeric(pCDR1CDR2),digits=3), x = unit(0.25, "npc"),y = unit(0.75, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) - grid.text(formatC(as.numeric(pFWR1CDR2),digits=3), x = unit(0.25, "npc"),y = unit(0.25, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) - - - # grid.text(paste("P = ",formatC(pCDRFWR,digits=3)), x = unit(0.5, "npc"),y = unit(0.98, "npc"),just=c("center", "top"),gp = gpar(cex=cex)) - # grid.text(paste("P = ",formatC(pFWRFWR,digits=3)), x = unit(0.5, "npc"),y = unit(0.02, "npc"),just=c("center", "bottom"),gp = gpar(cex=cex)) - } - else{ - } -} - - -################################################################################## -################## The whole OCD's matrix ######################################## -################################################################################## - -#pdf(width=4*numbSeqs+1/3,height=4*numbSeqs+1/3) -pdf( output ,width=4*numbSeqs+1/3,height=4*numbSeqs+1/3) - -pushViewport(viewport(x=0.02,y=0.02,just = c("left", "bottom"),w =0.96,height=0.96,layout = grid.layout(numbSeqs+1,numbSeqs+1,widths=unit.c(unit(rep(1,numbSeqs),"null"),unit(4,"lines")),heights=unit.c(unit(4,"lines"),unit(rep(1,numbSeqs),"null"))))) - -for( seqOne in 1:numbSeqs+1){ - pushViewport(viewport(layout.pos.col = seqOne-1, layout.pos.row = 1)) - if(seqOne>2){ - grid.polygon(c(0,0,0.5,0.5),c(0,0.5,0.5,0),gp=gpar(col=grey(0.5),fill=grey(0.9)),default.units="npc") - grid.polygon(c(1,1,0.5,0.5),c(0,0.5,0.5,0),gp=gpar(col=grey(0.5),fill=grey(0.95)),default.units="npc") - grid.polygon(c(0,0,1,1),c(1,0.5,0.5,1),gp=gpar(col=grey(0.5)),default.units="npc") - - grid.text(y=.25,x=0.75,"FWR",gp = gpar(cex=1.5),just="center") - grid.text(y=.25,x=0.25,"CDR",gp = gpar(cex=1.5),just="center") - } - grid.rect(gp = gpar(col=grey(0.9))) - grid.text(y=.75,substr(paste(names(listPDFs)[rowIDs[seqOne-1]]),1,16),gp = gpar(cex=2),just="center") - popViewport(1) -} - -for( seqOne in 1:numbSeqs+1){ - pushViewport(viewport(layout.pos.row = seqOne, layout.pos.col = numbSeqs+1)) - if(seqOne<=numbSeqs){ - grid.polygon(c(0,0.5,0.5,0),c(0,0,0.5,0.5),gp=gpar(col=grey(0.5),fill=grey(0.95)),default.units="npc") - grid.polygon(c(0,0.5,0.5,0),c(1,1,0.5,0.5),gp=gpar(col=grey(0.5),fill=grey(0.9)),default.units="npc") - grid.polygon(c(1,0.5,0.5,1),c(0,0,1,1),gp=gpar(col=grey(0.5)),default.units="npc") - grid.text(x=.25,y=0.75,"CDR",gp = gpar(cex=1.5),just="center",rot=270) - grid.text(x=.25,y=0.25,"FWR",gp = gpar(cex=1.5),just="center",rot=270) - } - grid.rect(gp = gpar(col=grey(0.9))) - grid.text(x=0.75,substr(paste(names(listPDFs)[rowIDs[seqOne-1]]),1,16),gp = gpar(cex=2),rot=270,just="center") - popViewport(1) -} - -for( seqOne in 1:numbSeqs+1){ - for(seqTwo in 1:numbSeqs+1){ - pushViewport(viewport(layout.pos.col = seqTwo-1, layout.pos.row = seqOne)) - if(seqTwo>seqOne){ - plot_pvals(rowIDs[seqOne-1],rowIDs[seqTwo-1],cex=2) - grid.rect() - } - popViewport(1) - } -} - - -xMin=0 -xMax=0.01 -for(pdf1 in rowIDs){ - xMin_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][1] - xMin_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][1] - xMax_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001])] - xMax_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001])] - xMin=min(c(xMin_CDR,xMin_FWR,xMin),na.rm=TRUE) - xMax=max(c(xMax_CDR,xMax_FWR,xMax),na.rm=TRUE) -} - - - -for(i in 1:numbSeqs+1){ - for(j in (i-1):numbSeqs){ - pushViewport(viewport(layout.pos.col = i-1, layout.pos.row = j+1)) - grid.rect() - plot_grid_s(rowIDs[i-1],rowIDs[j],cex=1) - popViewport(1) - } -} - -dev.off() - -cat("Success", paste(rowIDs,collapse="_"),sep=":") - +options("warn"=-1) + +#from http://selection.med.yale.edu/baseline/Archive/Baseline%20Version%201.3/Baseline_Functions_Version1.3.r +# Compute p-value of two distributions +compareTwoDistsFaster <-function(sigma_S=seq(-20,20,length.out=4001), N=10000, dens1=runif(4001,0,1), dens2=runif(4001,0,1)){ +#print(c(length(dens1),length(dens2))) +if(length(dens1)>1 & length(dens2)>1 ){ + dens1<-dens1/sum(dens1) + dens2<-dens2/sum(dens2) + cum2 <- cumsum(dens2)-dens2/2 + tmp<- sum(sapply(1:length(dens1),function(i)return(dens1[i]*cum2[i]))) + #print(tmp) + if(tmp>0.5)tmp<-tmp-1 + return( tmp ) + } + else { + return(NA) + } + #return (sum(sapply(1:N,function(i)(sample(sigma_S,1,prob=dens1)>sample(sigma_S,1,prob=dens2))))/N) +} + + +require("grid") +arg <- commandArgs(TRUE) +#arg <- c("300143","4","5") +arg[!arg=="clonal"] +input <- arg[1] +output <- arg[2] +rowIDs <- as.numeric( sapply(arg[3:(max(3,length(arg)))],function(x){ gsub("chkbx","",x) } ) ) + +numbSeqs = length(rowIDs) + +if ( is.na(rowIDs[1]) | numbSeqs>10 ) { + stop( paste("Error: Please select between one and 10 seqeunces to compare.") ) +} + +#load( paste("output/",sessionID,".RData",sep="") ) +load( input ) +#input + +xMarks = seq(-20,20,length.out=4001) + +plot_grid_s<-function(pdf1,pdf2,Sample=100,cex=1,xlim=NULL,xMarks = seq(-20,20,length.out=4001)){ + yMax = max(c(abs(as.numeric(unlist(listPDFs[pdf1]))),abs(as.numeric(unlist(listPDFs[pdf2]))),0),na.rm=T) * 1.1 + + if(length(xlim==2)){ + xMin=xlim[1] + xMax=xlim[2] + } else { + xMin_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][1] + xMin_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][1] + xMax_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001])] + xMax_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001])] + + xMin_CDR2 = xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001][1] + xMin_FWR2 = xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001][1] + xMax_CDR2 = xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf2][[1]][["CDR"]]>0.001])] + xMax_FWR2 = xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf2][[1]][["FWR"]]>0.001])] + + xMin=min(c(xMin_CDR,xMin_FWR,xMin_CDR2,xMin_FWR2,0),na.rm=TRUE) + xMax=max(c(xMax_CDR,xMax_FWR,xMax_CDR2,xMax_FWR2,0),na.rm=TRUE) + } + + sigma<-approx(xMarks,xout=seq(xMin,xMax,length.out=Sample))$x + grid.rect(gp = gpar(col=gray(0.6),fill="white",cex=cex)) + x <- sigma + pushViewport(viewport(x=0.175,y=0.175,width=0.825,height=0.825,just=c("left","bottom"),default.units="npc")) + #pushViewport(plotViewport(c(1.8, 1.8, 0.25, 0.25)*cex)) + pushViewport(dataViewport(x, c(yMax,-yMax),gp = gpar(cex=cex),extension=c(0.05))) + grid.polygon(c(0,0,1,1),c(0,0.5,0.5,0),gp=gpar(col=grey(0.95),fill=grey(0.95)),default.units="npc") + grid.polygon(c(0,0,1,1),c(1,0.5,0.5,1),gp=gpar(col=grey(0.9),fill=grey(0.9)),default.units="npc") + grid.rect() + grid.xaxis(gp = gpar(cex=cex/1.1)) + yticks = pretty(c(-yMax,yMax),8) + yticks = yticks[yticks>(-yMax) & yticks<(yMax)] + grid.yaxis(at=yticks,label=abs(yticks),gp = gpar(cex=cex/1.1)) + if(length(listPDFs[pdf1][[1]][["CDR"]])>1){ + ycdr<-approx(xMarks,listPDFs[pdf1][[1]][["CDR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y + grid.lines(unit(x,"native"), unit(ycdr,"native"),gp=gpar(col=2,lwd=2)) + } + if(length(listPDFs[pdf1][[1]][["FWR"]])>1){ + yfwr<-approx(xMarks,listPDFs[pdf1][[1]][["FWR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y + grid.lines(unit(x,"native"), unit(-yfwr,"native"),gp=gpar(col=4,lwd=2)) + } + + if(length(listPDFs[pdf2][[1]][["CDR"]])>1){ + ycdr2<-approx(xMarks,listPDFs[pdf2][[1]][["CDR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y + grid.lines(unit(x,"native"), unit(ycdr2,"native"),gp=gpar(col=2,lwd=2,lty=2)) + } + if(length(listPDFs[pdf2][[1]][["FWR"]])>1){ + yfwr2<-approx(xMarks,listPDFs[pdf2][[1]][["FWR"]],xout=seq(xMin,xMax,length.out=Sample),yleft=0,yright=0)$y + grid.lines(unit(x,"native"), unit(-yfwr2,"native"),gp=gpar(col=4,lwd=2,lty=2)) + } + + grid.lines(unit(c(0,1),"npc"), unit(c(0.5,0.5),"npc"),gp=gpar(col=1)) + grid.lines(unit(c(0,0),"native"), unit(c(0,1),"npc"),gp=gpar(col=1,lwd=1,lty=3)) + + grid.text("All", x = unit(-2.5, "lines"), rot = 90,gp = gpar(cex=cex)) + grid.text( expression(paste("Selection Strength (", Sigma, ")", sep="")) , y = unit(-2.5, "lines"),gp = gpar(cex=cex)) + + if(pdf1==pdf2 & length(listPDFs[pdf2][[1]][["FWR"]])>1 & length(listPDFs[pdf2][[1]][["CDR"]])>1 ){ + pCDRFWR = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["CDR"]], dens2=listPDFs[[pdf1]][["FWR"]]) + pval = formatC(as.numeric(pCDRFWR),digits=3) + grid.text( substitute(expression(paste(P[CDR/FWR], "=", x, sep="")),list(x=pval))[[2]] , x = unit(0.02, "npc"),y = unit(0.98, "npc"),just=c("left", "top"),gp = gpar(cex=cex*1.2)) + } + grid.text(paste("CDR"), x = unit(0.98, "npc"),y = unit(0.98, "npc"),just=c("right", "top"),gp = gpar(cex=cex*1.5)) + grid.text(paste("FWR"), x = unit(0.98, "npc"),y = unit(0.02, "npc"),just=c("right", "bottom"),gp = gpar(cex=cex*1.5)) + popViewport(2) +} +#plot_grid_s(1) + + +p2col<-function(p=0.01){ + breaks=c(-.51,-0.1,-.05,-0.01,-0.005,0,0.005,0.01,0.05,0.1,0.51) + i<-findInterval(p,breaks) + cols = c( rgb(0.8,1,0.8), rgb(0.6,1,0.6), rgb(0.4,1,0.4), rgb(0.2,1,0.2) , rgb(0,1,0), + rgb(1,0,0), rgb(1,.2,.2), rgb(1,.4,.4), rgb(1,.6,.6) , rgb(1,.8,.8) ) + return(cols[i]) +} + + +plot_pvals<-function(pdf1,pdf2,cex=1,upper=TRUE){ + if(upper){ + pCDR1FWR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["CDR"]], dens2=listPDFs[[pdf2]][["FWR"]]) + pFWR1FWR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens1=listPDFs[[pdf1]][["FWR"]], dens2=listPDFs[[pdf2]][["FWR"]]) + pFWR1CDR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens2=listPDFs[[pdf2]][["CDR"]], dens1=listPDFs[[pdf1]][["FWR"]]) + pCDR1CDR2 = compareTwoDistsFaster(sigma_S=xMarks, N=10000, dens2=listPDFs[[pdf2]][["CDR"]], dens1=listPDFs[[pdf1]][["CDR"]]) + grid.polygon(c(0.5,0.5,1,1),c(0,0.5,0.5,0),gp=gpar(col=p2col(pFWR1FWR2),fill=p2col(pFWR1FWR2)),default.units="npc") + grid.polygon(c(0.5,0.5,1,1),c(1,0.5,0.5,1),gp=gpar(col=p2col(pCDR1FWR2),fill=p2col(pCDR1FWR2)),default.units="npc") + grid.polygon(c(0.5,0.5,0,0),c(1,0.5,0.5,1),gp=gpar(col=p2col(pCDR1CDR2),fill=p2col(pCDR1CDR2)),default.units="npc") + grid.polygon(c(0.5,0.5,0,0),c(0,0.5,0.5,0),gp=gpar(col=p2col(pFWR1CDR2),fill=p2col(pFWR1CDR2)),default.units="npc") + + grid.lines(c(0,1),0.5,gp=gpar(lty=2,col=gray(0.925))) + grid.lines(0.5,c(0,1),gp=gpar(lty=2,col=gray(0.925))) + + grid.text(formatC(as.numeric(pFWR1FWR2),digits=3), x = unit(0.75, "npc"),y = unit(0.25, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) + grid.text(formatC(as.numeric(pCDR1FWR2),digits=3), x = unit(0.75, "npc"),y = unit(0.75, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) + grid.text(formatC(as.numeric(pCDR1CDR2),digits=3), x = unit(0.25, "npc"),y = unit(0.75, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) + grid.text(formatC(as.numeric(pFWR1CDR2),digits=3), x = unit(0.25, "npc"),y = unit(0.25, "npc"),just=c("center", "center"),gp = gpar(cex=cex)) + + + # grid.text(paste("P = ",formatC(pCDRFWR,digits=3)), x = unit(0.5, "npc"),y = unit(0.98, "npc"),just=c("center", "top"),gp = gpar(cex=cex)) + # grid.text(paste("P = ",formatC(pFWRFWR,digits=3)), x = unit(0.5, "npc"),y = unit(0.02, "npc"),just=c("center", "bottom"),gp = gpar(cex=cex)) + } + else{ + } +} + + +################################################################################## +################## The whole OCD's matrix ######################################## +################################################################################## + +#pdf(width=4*numbSeqs+1/3,height=4*numbSeqs+1/3) +pdf( output ,width=4*numbSeqs+1/3,height=4*numbSeqs+1/3) + +pushViewport(viewport(x=0.02,y=0.02,just = c("left", "bottom"),w =0.96,height=0.96,layout = grid.layout(numbSeqs+1,numbSeqs+1,widths=unit.c(unit(rep(1,numbSeqs),"null"),unit(4,"lines")),heights=unit.c(unit(4,"lines"),unit(rep(1,numbSeqs),"null"))))) + +for( seqOne in 1:numbSeqs+1){ + pushViewport(viewport(layout.pos.col = seqOne-1, layout.pos.row = 1)) + if(seqOne>2){ + grid.polygon(c(0,0,0.5,0.5),c(0,0.5,0.5,0),gp=gpar(col=grey(0.5),fill=grey(0.9)),default.units="npc") + grid.polygon(c(1,1,0.5,0.5),c(0,0.5,0.5,0),gp=gpar(col=grey(0.5),fill=grey(0.95)),default.units="npc") + grid.polygon(c(0,0,1,1),c(1,0.5,0.5,1),gp=gpar(col=grey(0.5)),default.units="npc") + + grid.text(y=.25,x=0.75,"FWR",gp = gpar(cex=1.5),just="center") + grid.text(y=.25,x=0.25,"CDR",gp = gpar(cex=1.5),just="center") + } + grid.rect(gp = gpar(col=grey(0.9))) + grid.text(y=.75,substr(paste(names(listPDFs)[rowIDs[seqOne-1]]),1,16),gp = gpar(cex=2),just="center") + popViewport(1) +} + +for( seqOne in 1:numbSeqs+1){ + pushViewport(viewport(layout.pos.row = seqOne, layout.pos.col = numbSeqs+1)) + if(seqOne<=numbSeqs){ + grid.polygon(c(0,0.5,0.5,0),c(0,0,0.5,0.5),gp=gpar(col=grey(0.5),fill=grey(0.95)),default.units="npc") + grid.polygon(c(0,0.5,0.5,0),c(1,1,0.5,0.5),gp=gpar(col=grey(0.5),fill=grey(0.9)),default.units="npc") + grid.polygon(c(1,0.5,0.5,1),c(0,0,1,1),gp=gpar(col=grey(0.5)),default.units="npc") + grid.text(x=.25,y=0.75,"CDR",gp = gpar(cex=1.5),just="center",rot=270) + grid.text(x=.25,y=0.25,"FWR",gp = gpar(cex=1.5),just="center",rot=270) + } + grid.rect(gp = gpar(col=grey(0.9))) + grid.text(x=0.75,substr(paste(names(listPDFs)[rowIDs[seqOne-1]]),1,16),gp = gpar(cex=2),rot=270,just="center") + popViewport(1) +} + +for( seqOne in 1:numbSeqs+1){ + for(seqTwo in 1:numbSeqs+1){ + pushViewport(viewport(layout.pos.col = seqTwo-1, layout.pos.row = seqOne)) + if(seqTwo>seqOne){ + plot_pvals(rowIDs[seqOne-1],rowIDs[seqTwo-1],cex=2) + grid.rect() + } + popViewport(1) + } +} + + +xMin=0 +xMax=0.01 +for(pdf1 in rowIDs){ + xMin_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][1] + xMin_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][1] + xMax_CDR = xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["CDR"]]>0.001])] + xMax_FWR = xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001][length(xMarks[listPDFs[pdf1][[1]][["FWR"]]>0.001])] + xMin=min(c(xMin_CDR,xMin_FWR,xMin),na.rm=TRUE) + xMax=max(c(xMax_CDR,xMax_FWR,xMax),na.rm=TRUE) +} + + + +for(i in 1:numbSeqs+1){ + for(j in (i-1):numbSeqs){ + pushViewport(viewport(layout.pos.col = i-1, layout.pos.row = j+1)) + grid.rect() + plot_grid_s(rowIDs[i-1],rowIDs[j],cex=1) + popViewport(1) + } +} + +dev.off() + +cat("Success", paste(rowIDs,collapse="_"),sep=":") + diff -r 43a1aa648537 -r ba33b94637ca baseline/script_imgt.py --- a/baseline/script_imgt.py Thu Dec 07 03:44:38 2017 -0500 +++ b/baseline/script_imgt.py Tue Jan 29 03:54:09 2019 -0500 @@ -1,86 +1,86 @@ -#import xlrd #avoid dep -import argparse -import re - -parser = argparse.ArgumentParser() -parser.add_argument("--input", help="Excel input file containing one or more sheets where column G has the gene annotation, H has the sequence id and J has the sequence") -parser.add_argument("--ref", help="Reference file") -parser.add_argument("--output", help="Output file") -parser.add_argument("--id", help="ID to be used at the '>>>' line in the output") - -args = parser.parse_args() - -print "script_imgt.py" -print "input:", args.input -print "ref:", args.ref -print "output:", args.output -print "id:", args.id - -refdic = dict() -with open(args.ref, 'rU') as ref: - currentSeq = "" - currentId = "" - for line in ref: - if line.startswith(">"): - if currentSeq is not "" and currentId is not "": - refdic[currentId[1:]] = currentSeq - currentId = line.rstrip() - currentSeq = "" - else: - currentSeq += line.rstrip() - refdic[currentId[1:]] = currentSeq - -print "Have", str(len(refdic)), "reference sequences" - -vPattern = [r"(IGHV[0-9]-[0-9ab]+-?[0-9]?D?\*\d{1,2})"]#, -# r"(TRBV[0-9]{1,2}-?[0-9]?-?[123]?)", -# r"(IGKV[0-3]D?-[0-9]{1,2})", -# r"(IGLV[0-9]-[0-9]{1,2})", -# r"(TRAV[0-9]{1,2}(-[1-46])?(/DV[45678])?)", -# r"(TRGV[234589])", -# r"(TRDV[1-3])"] - -#vPattern = re.compile(r"|".join(vPattern)) -vPattern = re.compile("|".join(vPattern)) - -def filterGene(s, pattern): - if type(s) is not str: - return None - res = pattern.search(s) - if res: - return res.group(0) - return None - - - -currentSeq = "" -currentId = "" -first=True -with open(args.input, 'r') as i: - with open(args.output, 'a') as o: - o.write(">>>" + args.id + "\n") - outputdic = dict() - for line in i: - if first: - first = False - continue - linesplt = line.split("\t") - ref = filterGene(linesplt[1], vPattern) - if not ref or not linesplt[2].rstrip(): - continue - if ref in outputdic: - outputdic[ref] += [(linesplt[0].replace(">", ""), linesplt[2].replace(">", "").rstrip())] - else: - outputdic[ref] = [(linesplt[0].replace(">", ""), linesplt[2].replace(">", "").rstrip())] - #print outputdic - - for k in outputdic.keys(): - if k in refdic: - o.write(">>" + k + "\n") - o.write(refdic[k] + "\n") - for seq in outputdic[k]: - #print seq - o.write(">" + seq[0] + "\n") - o.write(seq[1] + "\n") - else: - print k + " not in reference, skipping " + k +#import xlrd #avoid dep +import argparse +import re + +parser = argparse.ArgumentParser() +parser.add_argument("--input", help="Excel input file containing one or more sheets where column G has the gene annotation, H has the sequence id and J has the sequence") +parser.add_argument("--ref", help="Reference file") +parser.add_argument("--output", help="Output file") +parser.add_argument("--id", help="ID to be used at the '>>>' line in the output") + +args = parser.parse_args() + +print "script_imgt.py" +print "input:", args.input +print "ref:", args.ref +print "output:", args.output +print "id:", args.id + +refdic = dict() +with open(args.ref, 'rU') as ref: + currentSeq = "" + currentId = "" + for line in ref: + if line.startswith(">"): + if currentSeq is not "" and currentId is not "": + refdic[currentId[1:]] = currentSeq + currentId = line.rstrip() + currentSeq = "" + else: + currentSeq += line.rstrip() + refdic[currentId[1:]] = currentSeq + +print "Have", str(len(refdic)), "reference sequences" + +vPattern = [r"(IGHV[0-9]-[0-9ab]+-?[0-9]?D?\*\d{1,2})"]#, +# r"(TRBV[0-9]{1,2}-?[0-9]?-?[123]?)", +# r"(IGKV[0-3]D?-[0-9]{1,2})", +# r"(IGLV[0-9]-[0-9]{1,2})", +# r"(TRAV[0-9]{1,2}(-[1-46])?(/DV[45678])?)", +# r"(TRGV[234589])", +# r"(TRDV[1-3])"] + +#vPattern = re.compile(r"|".join(vPattern)) +vPattern = re.compile("|".join(vPattern)) + +def filterGene(s, pattern): + if type(s) is not str: + return None + res = pattern.search(s) + if res: + return res.group(0) + return None + + + +currentSeq = "" +currentId = "" +first=True +with open(args.input, 'r') as i: + with open(args.output, 'a') as o: + o.write(">>>" + args.id + "\n") + outputdic = dict() + for line in i: + if first: + first = False + continue + linesplt = line.split("\t") + ref = filterGene(linesplt[1], vPattern) + if not ref or not linesplt[2].rstrip(): + continue + if ref in outputdic: + outputdic[ref] += [(linesplt[0].replace(">", ""), linesplt[2].replace(">", "").rstrip())] + else: + outputdic[ref] = [(linesplt[0].replace(">", ""), linesplt[2].replace(">", "").rstrip())] + #print outputdic + + for k in outputdic.keys(): + if k in refdic: + o.write(">>" + k + "\n") + o.write(refdic[k] + "\n") + for seq in outputdic[k]: + #print seq + o.write(">" + seq[0] + "\n") + o.write(seq[1] + "\n") + else: + print k + " not in reference, skipping " + k diff -r 43a1aa648537 -r ba33b94637ca baseline/script_xlsx.py --- a/baseline/script_xlsx.py Thu Dec 07 03:44:38 2017 -0500 +++ b/baseline/script_xlsx.py Tue Jan 29 03:54:09 2019 -0500 @@ -1,58 +1,58 @@ -import xlrd -import argparse - -parser = argparse.ArgumentParser() -parser.add_argument("--input", help="Excel input file containing one or more sheets where column G has the gene annotation, H has the sequence id and J has the sequence") -parser.add_argument("--ref", help="Reference file") -parser.add_argument("--output", help="Output file") - -args = parser.parse_args() - -gene_column = 6 -id_column = 7 -seq_column = 8 -LETTERS = [x for x in "ABCDEFGHIJKLMNOPQRSTUVWXYZ"] - - -refdic = dict() -with open(args.ref, 'r') as ref: - currentSeq = "" - currentId = "" - for line in ref.readlines(): - if line[0] is ">": - if currentSeq is not "" and currentId is not "": - refdic[currentId[1:]] = currentSeq - currentId = line.rstrip() - currentSeq = "" - else: - currentSeq += line.rstrip() - refdic[currentId[1:]] = currentSeq - -currentSeq = "" -currentId = "" -with xlrd.open_workbook(args.input, 'r') as wb: - with open(args.output, 'a') as o: - for sheet in wb.sheets(): - if sheet.cell(1,gene_column).value.find("IGHV") < 0: - print "Genes not in column " + LETTERS[gene_column] + ", skipping sheet " + sheet.name - continue - o.write(">>>" + sheet.name + "\n") - outputdic = dict() - for rowindex in range(1, sheet.nrows): - ref = sheet.cell(rowindex, gene_column).value.replace(">", "") - if ref in outputdic: - outputdic[ref] += [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)] - else: - outputdic[ref] = [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)] - #print outputdic - - for k in outputdic.keys(): - if k in refdic: - o.write(">>" + k + "\n") - o.write(refdic[k] + "\n") - for seq in outputdic[k]: - #print seq - o.write(">" + seq[0] + "\n") - o.write(seq[1] + "\n") - else: - print k + " not in reference, skipping " + k +import xlrd +import argparse + +parser = argparse.ArgumentParser() +parser.add_argument("--input", help="Excel input file containing one or more sheets where column G has the gene annotation, H has the sequence id and J has the sequence") +parser.add_argument("--ref", help="Reference file") +parser.add_argument("--output", help="Output file") + +args = parser.parse_args() + +gene_column = 6 +id_column = 7 +seq_column = 8 +LETTERS = [x for x in "ABCDEFGHIJKLMNOPQRSTUVWXYZ"] + + +refdic = dict() +with open(args.ref, 'r') as ref: + currentSeq = "" + currentId = "" + for line in ref.readlines(): + if line[0] is ">": + if currentSeq is not "" and currentId is not "": + refdic[currentId[1:]] = currentSeq + currentId = line.rstrip() + currentSeq = "" + else: + currentSeq += line.rstrip() + refdic[currentId[1:]] = currentSeq + +currentSeq = "" +currentId = "" +with xlrd.open_workbook(args.input, 'r') as wb: + with open(args.output, 'a') as o: + for sheet in wb.sheets(): + if sheet.cell(1,gene_column).value.find("IGHV") < 0: + print "Genes not in column " + LETTERS[gene_column] + ", skipping sheet " + sheet.name + continue + o.write(">>>" + sheet.name + "\n") + outputdic = dict() + for rowindex in range(1, sheet.nrows): + ref = sheet.cell(rowindex, gene_column).value.replace(">", "") + if ref in outputdic: + outputdic[ref] += [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)] + else: + outputdic[ref] = [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)] + #print outputdic + + for k in outputdic.keys(): + if k in refdic: + o.write(">>" + k + "\n") + o.write(refdic[k] + "\n") + for seq in outputdic[k]: + #print seq + o.write(">" + seq[0] + "\n") + o.write(seq[1] + "\n") + else: + print k + " not in reference, skipping " + k diff -r 43a1aa648537 -r ba33b94637ca baseline/wrapper.sh diff -r 43a1aa648537 -r ba33b94637ca change_o/DefineClones.py diff -r 43a1aa648537 -r ba33b94637ca change_o/MakeDb.py diff -r 43a1aa648537 -r ba33b94637ca change_o/change_o_url.txt --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/change_o/change_o_url.txt Tue Jan 29 03:54:09 2019 -0500 @@ -0,0 +1,1 @@ +https://changeo.readthedocs.io/en/version-0.4.4/ \ No newline at end of file diff -r 43a1aa648537 -r ba33b94637ca change_o/define_clones.sh diff -r 43a1aa648537 -r ba33b94637ca change_o/makedb.sh diff -r 43a1aa648537 -r ba33b94637ca merge_and_filter.r --- a/merge_and_filter.r Thu Dec 07 03:44:38 2017 -0500 +++ b/merge_and_filter.r Tue Jan 29 03:54:09 2019 -0500 @@ -1,303 +1,303 @@ -args <- commandArgs(trailingOnly = TRUE) - - -summaryfile = args[1] -sequencesfile = args[2] -mutationanalysisfile = args[3] -mutationstatsfile = args[4] -hotspotsfile = args[5] -aafile = args[6] -gene_identification_file= args[7] -output = args[8] -before.unique.file = args[9] -unmatchedfile = args[10] -method=args[11] -functionality=args[12] -unique.type=args[13] -filter.unique=args[14] -filter.unique.count=as.numeric(args[15]) -class.filter=args[16] -empty.region.filter=args[17] - -print(paste("filter.unique.count:", filter.unique.count)) - -summ = read.table(summaryfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -sequences = read.table(sequencesfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -mutationanalysis = read.table(mutationanalysisfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -mutationstats = read.table(mutationstatsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -hotspots = read.table(hotspotsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -AAs = read.table(aafile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") -gene_identification = read.table(gene_identification_file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") - -fix_column_names = function(df){ - if("V.DOMAIN.Functionality" %in% names(df)){ - names(df)[names(df) == "V.DOMAIN.Functionality"] = "Functionality" - print("found V.DOMAIN.Functionality, changed") - } - if("V.DOMAIN.Functionality.comment" %in% names(df)){ - names(df)[names(df) == "V.DOMAIN.Functionality.comment"] = "Functionality.comment" - print("found V.DOMAIN.Functionality.comment, changed") - } - return(df) -} - -fix_non_unique_ids = function(df){ - df$Sequence.ID = paste(df$Sequence.ID, 1:nrow(df)) - return(df) -} - -summ = fix_column_names(summ) -sequences = fix_column_names(sequences) -mutationanalysis = fix_column_names(mutationanalysis) -mutationstats = fix_column_names(mutationstats) -hotspots = fix_column_names(hotspots) -AAs = fix_column_names(AAs) - -if(method == "blastn"){ - #"qseqid\tsseqid\tpident\tlength\tmismatch\tgapopen\tqstart\tqend\tsstart\tsend\tevalue\tbitscore" - gene_identification = gene_identification[!duplicated(gene_identification$qseqid),] - ref_length = data.frame(sseqid=c("ca1", "ca2", "cg1", "cg2", "cg3", "cg4", "cm"), ref.length=c(81,81,141,141,141,141,52)) - gene_identification = merge(gene_identification, ref_length, by="sseqid", all.x=T) - gene_identification$chunk_hit_percentage = (gene_identification$length / gene_identification$ref.length) * 100 - gene_identification = gene_identification[,c("qseqid", "chunk_hit_percentage", "pident", "qstart", "sseqid")] - colnames(gene_identification) = c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match") -} - -#print("Summary analysis files columns") -#print(names(summ)) - - - -input.sequence.count = nrow(summ) -print(paste("Number of sequences in summary file:", input.sequence.count)) - -filtering.steps = data.frame(character(0), numeric(0)) - -filtering.steps = rbind(filtering.steps, c("Input", input.sequence.count)) - -filtering.steps[,1] = as.character(filtering.steps[,1]) -filtering.steps[,2] = as.character(filtering.steps[,2]) -#filtering.steps[,3] = as.numeric(filtering.steps[,3]) - -#print("summary files columns") -#print(names(summ)) - -summ = merge(summ, gene_identification, by="Sequence.ID") - -print(paste("Number of sequences after merging with gene identification:", nrow(summ))) - -summ = summ[summ$Functionality != "No results",] - -print(paste("Number of sequences after 'No results' filter:", nrow(summ))) - -filtering.steps = rbind(filtering.steps, c("After 'No results' filter", nrow(summ))) - -if(functionality == "productive"){ - summ = summ[summ$Functionality == "productive (see comment)" | summ$Functionality == "productive",] -} else if (functionality == "unproductive"){ - summ = summ[summ$Functionality == "unproductive (see comment)" | summ$Functionality == "unproductive",] -} else if (functionality == "remove_unknown"){ - summ = summ[summ$Functionality != "No results" & summ$Functionality != "unknown (see comment)" & summ$Functionality != "unknown",] -} - -print(paste("Number of sequences after functionality filter:", nrow(summ))) - -filtering.steps = rbind(filtering.steps, c("After functionality filter", nrow(summ))) - -if(F){ #to speed up debugging - set.seed(1) - summ = summ[sample(nrow(summ), floor(nrow(summ) * 0.03)),] - print(paste("Number of sequences after sampling 3%:", nrow(summ))) - - filtering.steps = rbind(filtering.steps, c("Number of sequences after sampling 3%", nrow(summ))) -} - -print("mutation analysis files columns") -print(names(mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])])) - -result = merge(summ, mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])], by="Sequence.ID") - -print(paste("Number of sequences after merging with mutation analysis file:", nrow(result))) - -#print("mutation stats files columns") -#print(names(mutationstats[,!(names(mutationstats) %in% names(result)[-1])])) - -result = merge(result, mutationstats[,!(names(mutationstats) %in% names(result)[-1])], by="Sequence.ID") - -print(paste("Number of sequences after merging with mutation stats file:", nrow(result))) - -print("hotspots files columns") -print(names(hotspots[,!(names(hotspots) %in% names(result)[-1])])) - -result = merge(result, hotspots[,!(names(hotspots) %in% names(result)[-1])], by="Sequence.ID") - -print(paste("Number of sequences after merging with hotspots file:", nrow(result))) - -print("sequences files columns") -print(c("FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT")) - -sequences = sequences[,c("Sequence.ID", "FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT")] -names(sequences) = c("Sequence.ID", "FR1.IMGT.seq", "CDR1.IMGT.seq", "FR2.IMGT.seq", "CDR2.IMGT.seq", "FR3.IMGT.seq", "CDR3.IMGT.seq") -result = merge(result, sequences, by="Sequence.ID", all.x=T) - -print("sequences files columns") -print("CDR3.IMGT") - -AAs = AAs[,c("Sequence.ID", "CDR3.IMGT")] -names(AAs) = c("Sequence.ID", "CDR3.IMGT.AA") -result = merge(result, AAs, by="Sequence.ID", all.x=T) - -print(paste("Number of sequences in result after merging with sequences:", nrow(result))) - -result$VGene = gsub("^Homsap ", "", result$V.GENE.and.allele) -result$VGene = gsub("[*].*", "", result$VGene) -result$DGene = gsub("^Homsap ", "", result$D.GENE.and.allele) -result$DGene = gsub("[*].*", "", result$DGene) -result$JGene = gsub("^Homsap ", "", result$J.GENE.and.allele) -result$JGene = gsub("[*].*", "", result$JGene) - -splt = strsplit(class.filter, "_")[[1]] -chunk_hit_threshold = as.numeric(splt[1]) -nt_hit_threshold = as.numeric(splt[2]) - -higher_than=(result$chunk_hit_percentage >= chunk_hit_threshold & result$nt_hit_percentage >= nt_hit_threshold) - -if(!all(higher_than, na.rm=T)){ #check for no unmatched - result[!higher_than,"best_match"] = paste("unmatched,", result[!higher_than,"best_match"]) -} - -if(class.filter == "101_101"){ - result$best_match = "all" -} - -write.table(x=result, file=gsub("merged.txt$", "before_filters.txt", output), sep="\t",quote=F,row.names=F,col.names=T) - -print(paste("Number of empty CDR1 sequences:", sum(result$CDR1.IMGT.seq == "", na.rm=T))) -print(paste("Number of empty FR2 sequences:", sum(result$FR2.IMGT.seq == "", na.rm=T))) -print(paste("Number of empty CDR2 sequences:", sum(result$CDR2.IMGT.seq == "", na.rm=T))) -print(paste("Number of empty FR3 sequences:", sum(result$FR3.IMGT.seq == "", na.rm=T))) - -if(empty.region.filter == "leader"){ - result = result[result$FR1.IMGT.seq != "" & result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] -} else if(empty.region.filter == "FR1"){ - result = result[result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] -} else if(empty.region.filter == "CDR1"){ - result = result[result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] -} else if(empty.region.filter == "FR2"){ - result = result[result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] -} - -print(paste("After removal sequences that are missing a gene region:", nrow(result))) -filtering.steps = rbind(filtering.steps, c("After removal sequences that are missing a gene region", nrow(result))) - -if(empty.region.filter == "leader"){ - result = result[!(grepl("n|N", result$FR1.IMGT.seq) | grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] -} else if(empty.region.filter == "FR1"){ - result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] -} else if(empty.region.filter == "CDR1"){ - result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] -} else if(empty.region.filter == "FR2"){ - result = result[!(grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] -} - -print(paste("Number of sequences in result after n filtering:", nrow(result))) -filtering.steps = rbind(filtering.steps, c("After N filter", nrow(result))) - -cleanup_columns = c("FR1.IMGT.Nb.of.mutations", - "CDR1.IMGT.Nb.of.mutations", - "FR2.IMGT.Nb.of.mutations", - "CDR2.IMGT.Nb.of.mutations", - "FR3.IMGT.Nb.of.mutations") - -for(col in cleanup_columns){ - result[,col] = gsub("\\(.*\\)", "", result[,col]) - result[,col] = as.numeric(result[,col]) - result[is.na(result[,col]),] = 0 -} - -write.table(result, before.unique.file, sep="\t", quote=F,row.names=F,col.names=T) - -if(filter.unique != "no"){ - clmns = names(result) - if(filter.unique == "remove_vjaa"){ - result$unique.def = paste(result$VGene, result$JGene, result$CDR3.IMGT.AA) - } else if(empty.region.filter == "leader"){ - result$unique.def = paste(result$FR1.IMGT.seq, result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) - } else if(empty.region.filter == "FR1"){ - result$unique.def = paste(result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) - } else if(empty.region.filter == "CDR1"){ - result$unique.def = paste(result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) - } else if(empty.region.filter == "FR2"){ - result$unique.def = paste(result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) - } - - if(grepl("remove", filter.unique)){ - result = result[duplicated(result$unique.def) | duplicated(result$unique.def, fromLast=T),] - unique.defs = data.frame(table(result$unique.def)) - unique.defs = unique.defs[unique.defs$Freq >= filter.unique.count,] - result = result[result$unique.def %in% unique.defs$Var1,] - } - - if(filter.unique != "remove_vjaa"){ - result$unique.def = paste(result$unique.def, gsub(",.*", "", result$best_match)) #keep the unique sequences that are in multiple classes, gsub so the unmatched don't have a class after it - } - - result = result[!duplicated(result$unique.def),] -} - -write.table(result, gsub("before_unique_filter.txt", "after_unique_filter.txt", before.unique.file), sep="\t", quote=F,row.names=F,col.names=T) - -filtering.steps = rbind(filtering.steps, c("After filter unique sequences", nrow(result))) - -print(paste("Number of sequences in result after unique filtering:", nrow(result))) - -if(nrow(summ) == 0){ - stop("No data remaining after filter") -} - -result$best_match_class = gsub(",.*", "", result$best_match) #gsub so the unmatched don't have a class after it - -#result$past = "" -#cls = unlist(strsplit(unique.type, ",")) -#for (i in 1:nrow(result)){ -# result[i,"past"] = paste(result[i,cls], collapse=":") -#} - - - -result$past = do.call(paste, c(result[unlist(strsplit(unique.type, ","))], sep = ":")) - -result.matched = result[!grepl("unmatched", result$best_match),] -result.unmatched = result[grepl("unmatched", result$best_match),] - -result = rbind(result.matched, result.unmatched) - -result = result[!(duplicated(result$past)), ] - -result = result[,!(names(result) %in% c("past", "best_match_class"))] - -print(paste("Number of sequences in result after", unique.type, "filtering:", nrow(result))) - -filtering.steps = rbind(filtering.steps, c("After remove duplicates based on filter", nrow(result))) - -unmatched = result[grepl("^unmatched", result$best_match),c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match")] - -print(paste("Number of rows in result:", nrow(result))) -print(paste("Number of rows in unmatched:", nrow(unmatched))) - -matched.sequences = result[!grepl("^unmatched", result$best_match),] - -write.table(x=matched.sequences, file=gsub("merged.txt$", "filtered.txt", output), sep="\t",quote=F,row.names=F,col.names=T) - -matched.sequences.count = nrow(matched.sequences) -unmatched.sequences.count = sum(grepl("^unmatched", result$best_match)) - -filtering.steps = rbind(filtering.steps, c("Number of matched sequences", matched.sequences.count)) -filtering.steps = rbind(filtering.steps, c("Number of unmatched sequences", unmatched.sequences.count)) -filtering.steps[,2] = as.numeric(filtering.steps[,2]) -filtering.steps$perc = round(filtering.steps[,2] / input.sequence.count * 100, 2) - -write.table(x=filtering.steps, file=gsub("unmatched", "filtering_steps", unmatchedfile), sep="\t",quote=F,row.names=F,col.names=F) - -write.table(x=result, file=output, sep="\t",quote=F,row.names=F,col.names=T) -write.table(x=unmatched, file=unmatchedfile, sep="\t",quote=F,row.names=F,col.names=T) +args <- commandArgs(trailingOnly = TRUE) + + +summaryfile = args[1] +sequencesfile = args[2] +mutationanalysisfile = args[3] +mutationstatsfile = args[4] +hotspotsfile = args[5] +aafile = args[6] +gene_identification_file= args[7] +output = args[8] +before.unique.file = args[9] +unmatchedfile = args[10] +method=args[11] +functionality=args[12] +unique.type=args[13] +filter.unique=args[14] +filter.unique.count=as.numeric(args[15]) +class.filter=args[16] +empty.region.filter=args[17] + +print(paste("filter.unique.count:", filter.unique.count)) + +summ = read.table(summaryfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +sequences = read.table(sequencesfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +mutationanalysis = read.table(mutationanalysisfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +mutationstats = read.table(mutationstatsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +hotspots = read.table(hotspotsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +AAs = read.table(aafile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") +gene_identification = read.table(gene_identification_file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="") + +fix_column_names = function(df){ + if("V.DOMAIN.Functionality" %in% names(df)){ + names(df)[names(df) == "V.DOMAIN.Functionality"] = "Functionality" + print("found V.DOMAIN.Functionality, changed") + } + if("V.DOMAIN.Functionality.comment" %in% names(df)){ + names(df)[names(df) == "V.DOMAIN.Functionality.comment"] = "Functionality.comment" + print("found V.DOMAIN.Functionality.comment, changed") + } + return(df) +} + +fix_non_unique_ids = function(df){ + df$Sequence.ID = paste(df$Sequence.ID, 1:nrow(df)) + return(df) +} + +summ = fix_column_names(summ) +sequences = fix_column_names(sequences) +mutationanalysis = fix_column_names(mutationanalysis) +mutationstats = fix_column_names(mutationstats) +hotspots = fix_column_names(hotspots) +AAs = fix_column_names(AAs) + +if(method == "blastn"){ + #"qseqid\tsseqid\tpident\tlength\tmismatch\tgapopen\tqstart\tqend\tsstart\tsend\tevalue\tbitscore" + gene_identification = gene_identification[!duplicated(gene_identification$qseqid),] + ref_length = data.frame(sseqid=c("ca1", "ca2", "cg1", "cg2", "cg3", "cg4", "cm"), ref.length=c(81,81,141,141,141,141,52)) + gene_identification = merge(gene_identification, ref_length, by="sseqid", all.x=T) + gene_identification$chunk_hit_percentage = (gene_identification$length / gene_identification$ref.length) * 100 + gene_identification = gene_identification[,c("qseqid", "chunk_hit_percentage", "pident", "qstart", "sseqid")] + colnames(gene_identification) = c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match") +} + +#print("Summary analysis files columns") +#print(names(summ)) + + + +input.sequence.count = nrow(summ) +print(paste("Number of sequences in summary file:", input.sequence.count)) + +filtering.steps = data.frame(character(0), numeric(0)) + +filtering.steps = rbind(filtering.steps, c("Input", input.sequence.count)) + +filtering.steps[,1] = as.character(filtering.steps[,1]) +filtering.steps[,2] = as.character(filtering.steps[,2]) +#filtering.steps[,3] = as.numeric(filtering.steps[,3]) + +#print("summary files columns") +#print(names(summ)) + +summ = merge(summ, gene_identification, by="Sequence.ID") + +print(paste("Number of sequences after merging with gene identification:", nrow(summ))) + +summ = summ[summ$Functionality != "No results",] + +print(paste("Number of sequences after 'No results' filter:", nrow(summ))) + +filtering.steps = rbind(filtering.steps, c("After 'No results' filter", nrow(summ))) + +if(functionality == "productive"){ + summ = summ[summ$Functionality == "productive (see comment)" | summ$Functionality == "productive",] +} else if (functionality == "unproductive"){ + summ = summ[summ$Functionality == "unproductive (see comment)" | summ$Functionality == "unproductive",] +} else if (functionality == "remove_unknown"){ + summ = summ[summ$Functionality != "No results" & summ$Functionality != "unknown (see comment)" & summ$Functionality != "unknown",] +} + +print(paste("Number of sequences after functionality filter:", nrow(summ))) + +filtering.steps = rbind(filtering.steps, c("After functionality filter", nrow(summ))) + +if(F){ #to speed up debugging + set.seed(1) + summ = summ[sample(nrow(summ), floor(nrow(summ) * 0.03)),] + print(paste("Number of sequences after sampling 3%:", nrow(summ))) + + filtering.steps = rbind(filtering.steps, c("Number of sequences after sampling 3%", nrow(summ))) +} + +print("mutation analysis files columns") +print(names(mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])])) + +result = merge(summ, mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])], by="Sequence.ID") + +print(paste("Number of sequences after merging with mutation analysis file:", nrow(result))) + +#print("mutation stats files columns") +#print(names(mutationstats[,!(names(mutationstats) %in% names(result)[-1])])) + +result = merge(result, mutationstats[,!(names(mutationstats) %in% names(result)[-1])], by="Sequence.ID") + +print(paste("Number of sequences after merging with mutation stats file:", nrow(result))) + +print("hotspots files columns") +print(names(hotspots[,!(names(hotspots) %in% names(result)[-1])])) + +result = merge(result, hotspots[,!(names(hotspots) %in% names(result)[-1])], by="Sequence.ID") + +print(paste("Number of sequences after merging with hotspots file:", nrow(result))) + +print("sequences files columns") +print(c("FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT")) + +sequences = sequences[,c("Sequence.ID", "FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT")] +names(sequences) = c("Sequence.ID", "FR1.IMGT.seq", "CDR1.IMGT.seq", "FR2.IMGT.seq", "CDR2.IMGT.seq", "FR3.IMGT.seq", "CDR3.IMGT.seq") +result = merge(result, sequences, by="Sequence.ID", all.x=T) + +print("sequences files columns") +print("CDR3.IMGT") + +AAs = AAs[,c("Sequence.ID", "CDR3.IMGT")] +names(AAs) = c("Sequence.ID", "CDR3.IMGT.AA") +result = merge(result, AAs, by="Sequence.ID", all.x=T) + +print(paste("Number of sequences in result after merging with sequences:", nrow(result))) + +result$VGene = gsub("^Homsap ", "", result$V.GENE.and.allele) +result$VGene = gsub("[*].*", "", result$VGene) +result$DGene = gsub("^Homsap ", "", result$D.GENE.and.allele) +result$DGene = gsub("[*].*", "", result$DGene) +result$JGene = gsub("^Homsap ", "", result$J.GENE.and.allele) +result$JGene = gsub("[*].*", "", result$JGene) + +splt = strsplit(class.filter, "_")[[1]] +chunk_hit_threshold = as.numeric(splt[1]) +nt_hit_threshold = as.numeric(splt[2]) + +higher_than=(result$chunk_hit_percentage >= chunk_hit_threshold & result$nt_hit_percentage >= nt_hit_threshold) + +if(!all(higher_than, na.rm=T)){ #check for no unmatched + result[!higher_than,"best_match"] = paste("unmatched,", result[!higher_than,"best_match"]) +} + +if(class.filter == "101_101"){ + result$best_match = "all" +} + +write.table(x=result, file=gsub("merged.txt$", "before_filters.txt", output), sep="\t",quote=F,row.names=F,col.names=T) + +print(paste("Number of empty CDR1 sequences:", sum(result$CDR1.IMGT.seq == "", na.rm=T))) +print(paste("Number of empty FR2 sequences:", sum(result$FR2.IMGT.seq == "", na.rm=T))) +print(paste("Number of empty CDR2 sequences:", sum(result$CDR2.IMGT.seq == "", na.rm=T))) +print(paste("Number of empty FR3 sequences:", sum(result$FR3.IMGT.seq == "", na.rm=T))) + +if(empty.region.filter == "leader"){ + result = result[result$FR1.IMGT.seq != "" & result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] +} else if(empty.region.filter == "FR1"){ + result = result[result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] +} else if(empty.region.filter == "CDR1"){ + result = result[result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] +} else if(empty.region.filter == "FR2"){ + result = result[result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ] +} + +print(paste("After removal sequences that are missing a gene region:", nrow(result))) +filtering.steps = rbind(filtering.steps, c("After removal sequences that are missing a gene region", nrow(result))) + +if(empty.region.filter == "leader"){ + result = result[!(grepl("n|N", result$FR1.IMGT.seq) | grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] +} else if(empty.region.filter == "FR1"){ + result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] +} else if(empty.region.filter == "CDR1"){ + result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] +} else if(empty.region.filter == "FR2"){ + result = result[!(grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),] +} + +print(paste("Number of sequences in result after n filtering:", nrow(result))) +filtering.steps = rbind(filtering.steps, c("After N filter", nrow(result))) + +cleanup_columns = c("FR1.IMGT.Nb.of.mutations", + "CDR1.IMGT.Nb.of.mutations", + "FR2.IMGT.Nb.of.mutations", + "CDR2.IMGT.Nb.of.mutations", + "FR3.IMGT.Nb.of.mutations") + +for(col in cleanup_columns){ + result[,col] = gsub("\\(.*\\)", "", result[,col]) + result[,col] = as.numeric(result[,col]) + result[is.na(result[,col]),] = 0 +} + +write.table(result, before.unique.file, sep="\t", quote=F,row.names=F,col.names=T) + +if(filter.unique != "no"){ + clmns = names(result) + if(filter.unique == "remove_vjaa"){ + result$unique.def = paste(result$VGene, result$JGene, result$CDR3.IMGT.AA) + } else if(empty.region.filter == "leader"){ + result$unique.def = paste(result$FR1.IMGT.seq, result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) + } else if(empty.region.filter == "FR1"){ + result$unique.def = paste(result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) + } else if(empty.region.filter == "CDR1"){ + result$unique.def = paste(result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) + } else if(empty.region.filter == "FR2"){ + result$unique.def = paste(result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq) + } + + if(grepl("remove", filter.unique)){ + result = result[duplicated(result$unique.def) | duplicated(result$unique.def, fromLast=T),] + unique.defs = data.frame(table(result$unique.def)) + unique.defs = unique.defs[unique.defs$Freq >= filter.unique.count,] + result = result[result$unique.def %in% unique.defs$Var1,] + } + + if(filter.unique != "remove_vjaa"){ + result$unique.def = paste(result$unique.def, gsub(",.*", "", result$best_match)) #keep the unique sequences that are in multiple classes, gsub so the unmatched don't have a class after it + } + + result = result[!duplicated(result$unique.def),] +} + +write.table(result, gsub("before_unique_filter.txt", "after_unique_filter.txt", before.unique.file), sep="\t", quote=F,row.names=F,col.names=T) + +filtering.steps = rbind(filtering.steps, c("After filter unique sequences", nrow(result))) + +print(paste("Number of sequences in result after unique filtering:", nrow(result))) + +if(nrow(summ) == 0){ + stop("No data remaining after filter") +} + +result$best_match_class = gsub(",.*", "", result$best_match) #gsub so the unmatched don't have a class after it + +#result$past = "" +#cls = unlist(strsplit(unique.type, ",")) +#for (i in 1:nrow(result)){ +# result[i,"past"] = paste(result[i,cls], collapse=":") +#} + + + +result$past = do.call(paste, c(result[unlist(strsplit(unique.type, ","))], sep = ":")) + +result.matched = result[!grepl("unmatched", result$best_match),] +result.unmatched = result[grepl("unmatched", result$best_match),] + +result = rbind(result.matched, result.unmatched) + +result = result[!(duplicated(result$past)), ] + +result = result[,!(names(result) %in% c("past", "best_match_class"))] + +print(paste("Number of sequences in result after", unique.type, "filtering:", nrow(result))) + +filtering.steps = rbind(filtering.steps, c("After remove duplicates based on filter", nrow(result))) + +unmatched = result[grepl("^unmatched", result$best_match),c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match")] + +print(paste("Number of rows in result:", nrow(result))) +print(paste("Number of rows in unmatched:", nrow(unmatched))) + +matched.sequences = result[!grepl("^unmatched", result$best_match),] + +write.table(x=matched.sequences, file=gsub("merged.txt$", "filtered.txt", output), sep="\t",quote=F,row.names=F,col.names=T) + +matched.sequences.count = nrow(matched.sequences) +unmatched.sequences.count = sum(grepl("^unmatched", result$best_match)) + +filtering.steps = rbind(filtering.steps, c("Number of matched sequences", matched.sequences.count)) +filtering.steps = rbind(filtering.steps, c("Number of unmatched sequences", unmatched.sequences.count)) +filtering.steps[,2] = as.numeric(filtering.steps[,2]) +filtering.steps$perc = round(filtering.steps[,2] / input.sequence.count * 100, 2) + +write.table(x=filtering.steps, file=gsub("unmatched", "filtering_steps", unmatchedfile), sep="\t",quote=F,row.names=F,col.names=F) + +write.table(x=result, file=output, sep="\t",quote=F,row.names=F,col.names=T) +write.table(x=unmatched, file=unmatchedfile, sep="\t",quote=F,row.names=F,col.names=T) diff -r 43a1aa648537 -r ba33b94637ca shm_clonality.htm --- a/shm_clonality.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_clonality.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,144 +1,144 @@ - - - - - - - - - - - -
- -

References

- -

Gupta, -Namita T. and Vander Heiden, Jason A. and Uduman, Mohamed and Gadala-Maria, -Daniel and Yaari, Gur and Kleinstein, Steven H. (2015). Change-O: a toolkit for analyzing large-scale B cell -immunoglobulin repertoire sequencing data: Table 1. In Bioinformatics, 31 (20), pp. -3356–3358. [doi:10.1093/bioinformatics/btv359][Link]

- -

 

- -

All, IGA, IGG, IGM and IGE tabs

- -

In -these tabs information on the clonal relation of transcripts can be found. To -calculate clonal relation Change-O is used (Gupta et al, PMID: 26069265). -Transcripts are considered clonally related if they have maximal three nucleotides -difference in their CDR3 sequence and the same first V segment (as assigned by -IMGT). Results are represented in a table format showing the clone size and the -number of clones or sequences with this clone size. Change-O settings used are -the nucleotide hamming distance substitution model with -a complete distance of maximal three. For clonal assignment the first gene -segments were used, and the distances were not normalized. In case of -asymmetric distances, the minimal distance was used.

- -

 

- -

Overlap -tab

- -

This -tab gives information on with which (sub)classe(s) each unique analyzed region -(based on the exact nucleotide sequence of the analyzes region and the CDR3 -nucleotide sequence) is found with. This gives information if the combination -of the exact same nucleotide sequence of the analyzed region and the CDR3 -sequence can be found in multiple (sub)classes.

- -

Please note that this tab is based on all -sequences before filter unique sequences and the remove duplicates based on -filters are applied. In this table only sequences occuring more than once are -included.

- -
- - - - + + + + + + + + + + + +
+ +

References

+ +

Gupta, +Namita T. and Vander Heiden, Jason A. and Uduman, Mohamed and Gadala-Maria, +Daniel and Yaari, Gur and Kleinstein, Steven H. (2015). Change-O: a toolkit for analyzing large-scale B cell +immunoglobulin repertoire sequencing data: Table 1. In Bioinformatics, 31 (20), pp. +3356–3358. [doi:10.1093/bioinformatics/btv359][Link]

+ +

 

+ +

All, IGA, IGG, IGM and IGE tabs

+ +

In +these tabs information on the clonal relation of transcripts can be found. To +calculate clonal relation Change-O is used (Gupta et al, PMID: 26069265). +Transcripts are considered clonally related if they have maximal three nucleotides +difference in their CDR3 sequence and the same first V segment (as assigned by +IMGT). Results are represented in a table format showing the clone size and the +number of clones or sequences with this clone size. Change-O settings used are +the nucleotide hamming distance substitution model with +a complete distance of maximal three. For clonal assignment the first gene +segments were used, and the distances were not normalized. In case of +asymmetric distances, the minimal distance was used.

+ +

 

+ +

Overlap +tab

+ +

This +tab gives information on with which (sub)classe(s) each unique analyzed region +(based on the exact nucleotide sequence of the analyzes region and the CDR3 +nucleotide sequence) is found with. This gives information if the combination +of the exact same nucleotide sequence of the analyzed region and the CDR3 +sequence can be found in multiple (sub)classes.

+ +

Please note that this tab is based on all +sequences before filter unique sequences and the remove duplicates based on +filters are applied. In this table only sequences occuring more than once are +included.

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_csr.htm --- a/shm_csr.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_csr.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,95 +1,95 @@ - - - - - - - - - - - -
- -

The -graphs in this tab give insight into the subclass distribution of IGG and IGA -transcripts. Human Cµ, Cα, Cγ and Cε -constant genes are assigned using a custom script -specifically designed for human (sub)class assignment in repertoire data as -described in van Schouwenburg and IJspeert et al, submitted for publication. In -this script the reference sequences for the subclasses are divided in 8 -nucleotide chunks which overlap by 4 nucleotides. These overlapping chunks are -then individually aligned in the right order to each input sequence. The -percentage of the chunks identified in each rearrangement is calculated in the -‘chunk hit percentage’. Cα and Cγ -subclasses are very homologous and only differ in a few nucleotides. To assign -subclasses the ‘nt hit percentage’ is calculated. -This percentage indicates how well the chunks covering the subclass specific -nucleotide match with the different subclasses. Information -on normal distribution of subclasses in healthy individuals of different ages -can be found in IJspeert and van Schouwenburg et al, PMID: 27799928.

- -

IGA -subclass distribution

- -

Pie -chart showing the relative distribution of IGA1 and IGA2 transcripts in the -sample.

- -

IGG -subclass distribution

- -

Pie -chart showing the relative distribution of IGG1, IGG2, IGG3 and IGG4 -transcripts in the sample.

- -
- - - - + + + + + + + + + + + +
+ +

The +graphs in this tab give insight into the subclass distribution of IGG and IGA +transcripts. Human Cµ, Cα, Cγ and Cε +constant genes are assigned using a custom script +specifically designed for human (sub)class assignment in repertoire data as +described in van Schouwenburg and IJspeert et al, submitted for publication. In +this script the reference sequences for the subclasses are divided in 8 +nucleotide chunks which overlap by 4 nucleotides. These overlapping chunks are +then individually aligned in the right order to each input sequence. The +percentage of the chunks identified in each rearrangement is calculated in the +‘chunk hit percentage’. Cα and Cγ +subclasses are very homologous and only differ in a few nucleotides. To assign +subclasses the ‘nt hit percentage’ is calculated. +This percentage indicates how well the chunks covering the subclass specific +nucleotide match with the different subclasses. Information +on normal distribution of subclasses in healthy individuals of different ages +can be found in IJspeert and van Schouwenburg et al, PMID: 27799928.

+ +

IGA +subclass distribution

+ +

Pie +chart showing the relative distribution of IGA1 and IGA2 transcripts in the +sample.

+ +

IGG +subclass distribution

+ +

Pie +chart showing the relative distribution of IGG1, IGG2, IGG3 and IGG4 +transcripts in the sample.

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_csr.r diff -r 43a1aa648537 -r ba33b94637ca shm_csr.xml --- a/shm_csr.xml Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_csr.xml Tue Jan 29 03:54:09 2019 -0500 @@ -1,14 +1,13 @@ - + numpy + xlrd + r-ggplot2 + r-reshape2 + r-scales + r-seqinr + r-data.table #if str ( $filter_unique.filter_unique_select ) == "remove": diff -r 43a1aa648537 -r ba33b94637ca shm_downloads.htm --- a/shm_downloads.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_downloads.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,538 +1,538 @@ - - - - - - - - - - - -
- -

Info

- -

The complete -dataset: -Allows downloading of the complete parsed data set.

- -

The filtered -dataset: -Allows downloading of all parsed IMGT information of all transcripts that -passed the chosen filter settings.

- -

The alignment -info on the unmatched sequences: Provides information of the subclass -alignment of all unmatched sequences. For each sequence the chunck hit -percentage and the nt hit percentage is shown together with the best matched -subclass.

- -

SHM Overview

- -

The SHM Overview -table as a dataset: Allows downloading of the SHM Overview -table as a data set. 

- -

Motif data per -sequence ID: Provides a file that contains information for each -transcript on the number of mutations present in WA/TW and RGYW/WRCY motives.

- -

Mutation data -per sequence ID: Provides a file containing information -on the number of sequences bases, the number and location of mutations and the -type of mutations found in each transcript.

- -

Base count for -every sequence: links to a page showing for each transcript the -sequence of the analysed region (as dependent on the sequence starts at filter), -the assigned subclass and the number of sequenced A,C,G and T’s.

- -

The data used to -generate the percentage of mutations in AID and pol eta motives plot: -Provides a file containing the values used to generate the percentage of -mutations in AID and pol eta motives plot in the SHM overview tab.

- -

The -data used to generate the relative mutation patterns plot: -Provides a download with the data used to generate the relative mutation -patterns plot in the SHM overview tab.

- -

The -data used to generate the absolute mutation patterns plot: -Provides a download with the data used to generate the absolute mutation -patterns plot in the SHM overview tab.

- -

SHM Frequency

- -

The data -generate the frequency scatter plot: Allows -downloading the data used to generate the frequency scatter plot in the SHM -frequency tab.

- -

The data used to -generate the frequency by class plot: Allows -downloading the data used to generate frequency by class plot included in the -SHM frequency tab.          

- -

The data for -frequency by subclass: Provides information of the number and -percentage of sequences that have 0%, 0-2%, 2-5%, 5-10%, 10-15%, 15-20%, ->20% SHM. Information is provided for each subclass.

- -

 

- -

Transition -Tables

- -

The data for the -'all' transition plot: Contains the information used to -generate the transition table for all sequences.

- -

The data for the -'IGA' transition plot: Contains the information used to -generate the transition table for all IGA sequences.

- -

The data for the -'IGA1' transition plot: Contains the information used to -generate the transition table for all IGA1 sequences.

- -

The data for the -'IGA2' transition plot: Contains the information used to -generate the transition table for all IGA2 sequences.

- -

The data for the -'IGG' transition plot : Contains the information used to -generate the transition table for all IGG sequences.

- -

The data for the -'IGG1' transition plot: Contains the information used to -generate the transition table for all IGG1 sequences.

- -

The data for the -'IGG2' transition plot: Contains the information used to -generate the transition table for all IGG2 sequences.

- -

The data for the -'IGG3' transition plot: Contains the information used to -generate the transition table for all IGG3 sequences.

- -

The data for the -'IGG4' transition plot: Contains the information used to -generate the transition table for all IGG4 sequences.

- -

The data for the -'IGM' transition plot : Contains the information used to -generate the transition table for all IGM sequences.

- -

The data for the -'IGE' transition plot: Contains the -information used to generate the transition table for all IGE sequences.

- -

Antigen -selection

- -

AA mutation data -per sequence ID: Provides for each transcript information on whether -there is replacement mutation at each amino acid location (as defined by IMGT). -For all amino acids outside of the analysed region the value 0 is given.

- -

Presence of AA -per sequence ID: Provides for each transcript information on which -amino acid location (as defined by IMGT) is present. 0 is absent, 1 -is present.

- -

The data used to -generate the aa mutation frequency plot: Provides the -data used to generate the aa mutation frequency plot for all sequences in the -antigen selection tab.

- -

The data used to -generate the aa mutation frequency plot for IGA: Provides the -data used to generate the aa mutation frequency plot for all IGA sequences in -the antigen selection tab.

- -

The data used to -generate the aa mutation frequency plot for IGG: Provides the -data used to generate the aa mutation frequency plot for all IGG sequences in -the antigen selection tab.

- -

The data used to -generate the aa mutation frequency plot for IGM: Provides the -data used to generate the aa mutation frequency plot for all IGM sequences in -the antigen selection tab.

- -

The data used to -generate the aa mutation frequency plot for IGE:  Provides the -data used to generate the aa mutation frequency plot for all IGE sequences in -the antigen selection tab.

- -

Baseline PDF (http://selection.med.yale.edu/baseline/): PDF -containing the Antigen selection (BASELINe) graph for all -sequences.

- -

Baseline data: -Table output of the BASELINe analysis. Calculation of antigen selection as -performed by BASELINe are shown for each individual sequence and the sum of all -sequences.

- -

Baseline IGA -PDF: -PDF containing the Antigen selection (BASELINe) graph for all -sequences.

- -

Baseline IGA -data: -Table output of the BASELINe analysis. Calculation of antigen selection as -performed by BASELINe are shown for each individual IGA sequence and the sum of -all IGA sequences.

- -

Baseline IGG -PDF: -PDF containing the Antigen selection (BASELINe) graph for all IGG -sequences.

- -

Baseline IGG -data: -Table output of the BASELINe analysis. Calculation of antigen selection as -performed by BASELINe are shown for each individual IGG sequence and the sum of -all IGG sequences.       

- -

Baseline IGM PDF: PDF -containing the Antigen selection (BASELINe) graph for all IGM -sequences.

- -

Baseline IGM -data: -Table output of the BASELINe analysis. Calculation of antigen selection as -performed by BASELINe are shown for each individual IGM sequence and the sum of -all IGM sequences.

- -

Baseline IGE -PDF: -PDF containing the Antigen selection (BASELINe) graph for all IGE -sequences. -

- -

Baseline IGE -data: -Table output of the BASELINe analysis. Calculation of antigen selection as -performed by BASELINe are shown for each individual IGE sequence and the sum of -all IGE sequences.

- -

CSR

- -

The data for the -IGA -subclass distribution plot : Data used for -the generation of the IGA subclass distribution plot provided -in the CSR tab.

- -

The data for the -IGA -subclass distribution plot : Data used for the generation of the IGG -subclass distribution plot provided in the CSR tab.

- -

Clonal relation

- -

Sequence overlap -between subclasses: Link to the overlap table as provided -under the clonality overlap tab.        

- -

The Change-O DB -file with defined clones and subclass annotation: -Downloads a table with the calculation of clonal relation between all -sequences. For each individual transcript the results of the clonal assignment -as provided by Change-O are provided. Sequences with the same number in the CLONE -column are considered clonally related.

- -

The Change-O DB -defined clones summary file: Gives a summary of the total number of -clones in all sequences and their clone size.          

- -

The Change-O DB -file with defined clones of IGA: Downloads a table with the -calculation of clonal relation between all IGA sequences. For each individual -transcript the results of the clonal assignment as provided by Change-O are -provided. Sequences with the same number in the CLONE column are considered -clonally related.

- -

The Change-O DB -defined clones summary file of IGA: Gives a summary -of the total number of clones in all IGA sequences and their clone size.

- -

The Change-O DB -file with defined clones of IGG: Downloads a table with the -calculation of clonal relation between all IGG sequences. For each individual -transcript the results of the clonal assignment as provided by Change-O are -provided. Sequences with the same number in the CLONE column are considered -clonally related.

- -

The Change-O DB -defined clones summary file of IGG: Gives a summary -of the total number of clones in all IGG sequences and their clone size.

- -

The Change-O DB -file with defined clones of IGM: Downloads a table -with the calculation of clonal relation between all IGM sequences. For each -individual transcript the results of the clonal assignment as provided by -Change-O are provided. Sequences with the same number in the CLONE column are -considered clonally related.

- -

The Change-O DB -defined clones summary file of IGM: Gives a summary -of the total number of clones in all IGM sequences and their clone size.

- -

The Change-O DB -file with defined clones of IGE: Downloads a table with the -calculation of clonal relation between all IGE sequences. For each individual -transcript the results of the clonal assignment as provided by Change-O are -provided. Sequences with the same number in the CLONE column are considered -clonally related.

- -

The Change-O DB -defined clones summary file of IGE: Gives a summary -of the total number of clones in all IGE sequences and their clone size.

- -

Filtered IMGT -output files

- -

An IMGT archive -with just the matched and filtered sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGA sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGA -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGA1 sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGA1 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGA2 sequences: Downloads a .txz -file with the same format as downloaded IMGT files that contains all IGA2 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGG sequences: Downloads a .txz -file with the same format as downloaded IMGT files that contains all IGG -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGG1 sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGG1 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGG2 sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGG2 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGG3 sequences: Downloads a .txz -file with the same format as downloaded IMGT files that contains all IGG3 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGG4 sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGG4 -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGM sequences: Downloads a .txz -file with the same format as downloaded IMGT files that contains all IGM -sequences that have passed the chosen filter settings.

- -

An IMGT archive -with just the matched and filtered IGE sequences: Downloads a -.txz file with the same format as downloaded IMGT files that contains all IGE -sequences that have passed the chosen filter settings.

- -
- - - - + + + + + + + + + + + +
+ +

Info

+ +

The complete +dataset: +Allows downloading of the complete parsed data set.

+ +

The filtered +dataset: +Allows downloading of all parsed IMGT information of all transcripts that +passed the chosen filter settings.

+ +

The alignment +info on the unmatched sequences: Provides information of the subclass +alignment of all unmatched sequences. For each sequence the chunck hit +percentage and the nt hit percentage is shown together with the best matched +subclass.

+ +

SHM Overview

+ +

The SHM Overview +table as a dataset: Allows downloading of the SHM Overview +table as a data set. 

+ +

Motif data per +sequence ID: Provides a file that contains information for each +transcript on the number of mutations present in WA/TW and RGYW/WRCY motives.

+ +

Mutation data +per sequence ID: Provides a file containing information +on the number of sequences bases, the number and location of mutations and the +type of mutations found in each transcript.

+ +

Base count for +every sequence: links to a page showing for each transcript the +sequence of the analysed region (as dependent on the sequence starts at filter), +the assigned subclass and the number of sequenced A,C,G and T’s.

+ +

The data used to +generate the percentage of mutations in AID and pol eta motives plot: +Provides a file containing the values used to generate the percentage of +mutations in AID and pol eta motives plot in the SHM overview tab.

+ +

The +data used to generate the relative mutation patterns plot: +Provides a download with the data used to generate the relative mutation +patterns plot in the SHM overview tab.

+ +

The +data used to generate the absolute mutation patterns plot: +Provides a download with the data used to generate the absolute mutation +patterns plot in the SHM overview tab.

+ +

SHM Frequency

+ +

The data +generate the frequency scatter plot: Allows +downloading the data used to generate the frequency scatter plot in the SHM +frequency tab.

+ +

The data used to +generate the frequency by class plot: Allows +downloading the data used to generate frequency by class plot included in the +SHM frequency tab.          

+ +

The data for +frequency by subclass: Provides information of the number and +percentage of sequences that have 0%, 0-2%, 2-5%, 5-10%, 10-15%, 15-20%, +>20% SHM. Information is provided for each subclass.

+ +

 

+ +

Transition +Tables

+ +

The data for the +'all' transition plot: Contains the information used to +generate the transition table for all sequences.

+ +

The data for the +'IGA' transition plot: Contains the information used to +generate the transition table for all IGA sequences.

+ +

The data for the +'IGA1' transition plot: Contains the information used to +generate the transition table for all IGA1 sequences.

+ +

The data for the +'IGA2' transition plot: Contains the information used to +generate the transition table for all IGA2 sequences.

+ +

The data for the +'IGG' transition plot : Contains the information used to +generate the transition table for all IGG sequences.

+ +

The data for the +'IGG1' transition plot: Contains the information used to +generate the transition table for all IGG1 sequences.

+ +

The data for the +'IGG2' transition plot: Contains the information used to +generate the transition table for all IGG2 sequences.

+ +

The data for the +'IGG3' transition plot: Contains the information used to +generate the transition table for all IGG3 sequences.

+ +

The data for the +'IGG4' transition plot: Contains the information used to +generate the transition table for all IGG4 sequences.

+ +

The data for the +'IGM' transition plot : Contains the information used to +generate the transition table for all IGM sequences.

+ +

The data for the +'IGE' transition plot: Contains the +information used to generate the transition table for all IGE sequences.

+ +

Antigen +selection

+ +

AA mutation data +per sequence ID: Provides for each transcript information on whether +there is replacement mutation at each amino acid location (as defined by IMGT). +For all amino acids outside of the analysed region the value 0 is given.

+ +

Presence of AA +per sequence ID: Provides for each transcript information on which +amino acid location (as defined by IMGT) is present. 0 is absent, 1 +is present.

+ +

The data used to +generate the aa mutation frequency plot: Provides the +data used to generate the aa mutation frequency plot for all sequences in the +antigen selection tab.

+ +

The data used to +generate the aa mutation frequency plot for IGA: Provides the +data used to generate the aa mutation frequency plot for all IGA sequences in +the antigen selection tab.

+ +

The data used to +generate the aa mutation frequency plot for IGG: Provides the +data used to generate the aa mutation frequency plot for all IGG sequences in +the antigen selection tab.

+ +

The data used to +generate the aa mutation frequency plot for IGM: Provides the +data used to generate the aa mutation frequency plot for all IGM sequences in +the antigen selection tab.

+ +

The data used to +generate the aa mutation frequency plot for IGE:  Provides the +data used to generate the aa mutation frequency plot for all IGE sequences in +the antigen selection tab.

+ +

Baseline PDF (http://selection.med.yale.edu/baseline/): PDF +containing the Antigen selection (BASELINe) graph for all +sequences.

+ +

Baseline data: +Table output of the BASELINe analysis. Calculation of antigen selection as +performed by BASELINe are shown for each individual sequence and the sum of all +sequences.

+ +

Baseline IGA +PDF: +PDF containing the Antigen selection (BASELINe) graph for all +sequences.

+ +

Baseline IGA +data: +Table output of the BASELINe analysis. Calculation of antigen selection as +performed by BASELINe are shown for each individual IGA sequence and the sum of +all IGA sequences.

+ +

Baseline IGG +PDF: +PDF containing the Antigen selection (BASELINe) graph for all IGG +sequences.

+ +

Baseline IGG +data: +Table output of the BASELINe analysis. Calculation of antigen selection as +performed by BASELINe are shown for each individual IGG sequence and the sum of +all IGG sequences.       

+ +

Baseline IGM PDF: PDF +containing the Antigen selection (BASELINe) graph for all IGM +sequences.

+ +

Baseline IGM +data: +Table output of the BASELINe analysis. Calculation of antigen selection as +performed by BASELINe are shown for each individual IGM sequence and the sum of +all IGM sequences.

+ +

Baseline IGE +PDF: +PDF containing the Antigen selection (BASELINe) graph for all IGE +sequences. +

+ +

Baseline IGE +data: +Table output of the BASELINe analysis. Calculation of antigen selection as +performed by BASELINe are shown for each individual IGE sequence and the sum of +all IGE sequences.

+ +

CSR

+ +

The data for the +IGA +subclass distribution plot : Data used for +the generation of the IGA subclass distribution plot provided +in the CSR tab.

+ +

The data for the +IGA +subclass distribution plot : Data used for the generation of the IGG +subclass distribution plot provided in the CSR tab.

+ +

Clonal relation

+ +

Sequence overlap +between subclasses: Link to the overlap table as provided +under the clonality overlap tab.        

+ +

The Change-O DB +file with defined clones and subclass annotation: +Downloads a table with the calculation of clonal relation between all +sequences. For each individual transcript the results of the clonal assignment +as provided by Change-O are provided. Sequences with the same number in the CLONE +column are considered clonally related.

+ +

The Change-O DB +defined clones summary file: Gives a summary of the total number of +clones in all sequences and their clone size.          

+ +

The Change-O DB +file with defined clones of IGA: Downloads a table with the +calculation of clonal relation between all IGA sequences. For each individual +transcript the results of the clonal assignment as provided by Change-O are +provided. Sequences with the same number in the CLONE column are considered +clonally related.

+ +

The Change-O DB +defined clones summary file of IGA: Gives a summary +of the total number of clones in all IGA sequences and their clone size.

+ +

The Change-O DB +file with defined clones of IGG: Downloads a table with the +calculation of clonal relation between all IGG sequences. For each individual +transcript the results of the clonal assignment as provided by Change-O are +provided. Sequences with the same number in the CLONE column are considered +clonally related.

+ +

The Change-O DB +defined clones summary file of IGG: Gives a summary +of the total number of clones in all IGG sequences and their clone size.

+ +

The Change-O DB +file with defined clones of IGM: Downloads a table +with the calculation of clonal relation between all IGM sequences. For each +individual transcript the results of the clonal assignment as provided by +Change-O are provided. Sequences with the same number in the CLONE column are +considered clonally related.

+ +

The Change-O DB +defined clones summary file of IGM: Gives a summary +of the total number of clones in all IGM sequences and their clone size.

+ +

The Change-O DB +file with defined clones of IGE: Downloads a table with the +calculation of clonal relation between all IGE sequences. For each individual +transcript the results of the clonal assignment as provided by Change-O are +provided. Sequences with the same number in the CLONE column are considered +clonally related.

+ +

The Change-O DB +defined clones summary file of IGE: Gives a summary +of the total number of clones in all IGE sequences and their clone size.

+ +

Filtered IMGT +output files

+ +

An IMGT archive +with just the matched and filtered sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGA sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGA +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGA1 sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGA1 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGA2 sequences: Downloads a .txz +file with the same format as downloaded IMGT files that contains all IGA2 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGG sequences: Downloads a .txz +file with the same format as downloaded IMGT files that contains all IGG +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGG1 sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGG1 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGG2 sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGG2 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGG3 sequences: Downloads a .txz +file with the same format as downloaded IMGT files that contains all IGG3 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGG4 sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGG4 +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGM sequences: Downloads a .txz +file with the same format as downloaded IMGT files that contains all IGM +sequences that have passed the chosen filter settings.

+ +

An IMGT archive +with just the matched and filtered IGE sequences: Downloads a +.txz file with the same format as downloaded IMGT files that contains all IGE +sequences that have passed the chosen filter settings.

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_first.htm --- a/shm_first.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_first.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,127 +1,127 @@ - - - - - - - - - - - -
- -

Table showing the order of each -filtering step and the number and percentage of sequences after each filtering -step.

- -

Input: The -number of sequences in the original IMGT file. This is always 100% of the -sequences.

- -

After "no results" filter: IMGT -classifies sequences either as "productive", "unproductive", "unknown", or "no -results". Here, the number and percentages of sequences that are not classified -as "no results" are reported.

- -

After functionality filter: The -number and percentages of sequences that have passed the functionality filter. The -filtering performed is dependent on the settings of the functionality filter. -Details on the functionality filter can be found on the start page of -the SHM&CSR pipeline.

- -

After -removal sequences that are missing a gene region: -In this step all sequences that are missing a gene region (FR1, CDR1, FR2, -CDR2, FR3) that should be present are removed from analysis. The sequence -regions that should be present are dependent on the settings of the sequence -starts at filter. The number and -percentage of sequences that pass this filter step are reported.

- -

After -N filter: In this step all sequences that contain -an ambiguous base (n) in the analysed region or the CDR3 are removed from the -analysis. The analysed region is determined by the setting of the sequence -starts at filter. The number and percentage of sequences that pass this filter -step are reported.

- -

After -filter unique sequences: The number and -percentage of sequences that pass the "filter unique sequences" filter. Details -on this filter can be found on the start page of -the SHM&CSR pipeline

- -

After -remove duplicate based on filter: The number and -percentage of sequences that passed the remove duplicate filter. Details on the -"remove duplicate filter based on filter" can be found on the start page of the -SHM&CSR pipeline.

- -

Number of matches sequences: -The number and percentage of sequences that passed all the filters described -above and have a (sub)class assigned.

- -

Number -of unmatched sequences: The number and percentage -of sequences that passed all the filters described above and do not have -subclass assigned.

- -

 

- -
- - - - + + + + + + + + + + + +
+ +

Table showing the order of each +filtering step and the number and percentage of sequences after each filtering +step.

+ +

Input: The +number of sequences in the original IMGT file. This is always 100% of the +sequences.

+ +

After "no results" filter: IMGT +classifies sequences either as "productive", "unproductive", "unknown", or "no +results". Here, the number and percentages of sequences that are not classified +as "no results" are reported.

+ +

After functionality filter: The +number and percentages of sequences that have passed the functionality filter. The +filtering performed is dependent on the settings of the functionality filter. +Details on the functionality filter can be found on the start page of +the SHM&CSR pipeline.

+ +

After +removal sequences that are missing a gene region: +In this step all sequences that are missing a gene region (FR1, CDR1, FR2, +CDR2, FR3) that should be present are removed from analysis. The sequence +regions that should be present are dependent on the settings of the sequence +starts at filter. The number and +percentage of sequences that pass this filter step are reported.

+ +

After +N filter: In this step all sequences that contain +an ambiguous base (n) in the analysed region or the CDR3 are removed from the +analysis. The analysed region is determined by the setting of the sequence +starts at filter. The number and percentage of sequences that pass this filter +step are reported.

+ +

After +filter unique sequences: The number and +percentage of sequences that pass the "filter unique sequences" filter. Details +on this filter can be found on the start page of +the SHM&CSR pipeline

+ +

After +remove duplicate based on filter: The number and +percentage of sequences that passed the remove duplicate filter. Details on the +"remove duplicate filter based on filter" can be found on the start page of the +SHM&CSR pipeline.

+ +

Number of matches sequences: +The number and percentage of sequences that passed all the filters described +above and have a (sub)class assigned.

+ +

Number +of unmatched sequences: The number and percentage +of sequences that passed all the filters described above and do not have +subclass assigned.

+ +

 

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_frequency.htm --- a/shm_frequency.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_frequency.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,87 +1,87 @@ - - - - - - - - - - - -
- -

SHM -frequency tab

- -

Graphs

- -

These -graphs give insight into the level of SHM. The data represented in these graphs -can be downloaded in the download tab. More -information on the values found in healthy individuals of different ages can be -found in IJspeert and van Schouwenburg et al, PMID: 27799928.

- -

Frequency -scatter plot

- -

A -dot plot showing the percentage of SHM in each transcript divided into the -different (sub)classes. In the graph each dot -represents an individual transcript.

- -

Mutation -frequency by class

- -

A -bar graph showing the percentage of transcripts that contain 0%, 0-2%, 2-5%, -5-10% 10-15%, 15-20% or more than 20% SHM for each subclass.

- -

Hanna IJspeert, Pauline A. van -Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, -Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation -of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and -Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

- -
- - - - + + + + + + + + + + + +
+ +

SHM +frequency tab

+ +

Graphs

+ +

These +graphs give insight into the level of SHM. The data represented in these graphs +can be downloaded in the download tab. More +information on the values found in healthy individuals of different ages can be +found in IJspeert and van Schouwenburg et al, PMID: 27799928.

+ +

Frequency +scatter plot

+ +

A +dot plot showing the percentage of SHM in each transcript divided into the +different (sub)classes. In the graph each dot +represents an individual transcript.

+ +

Mutation +frequency by class

+ +

A +bar graph showing the percentage of transcripts that contain 0%, 0-2%, 2-5%, +5-10% 10-15%, 15-20% or more than 20% SHM for each subclass.

+ +

Hanna IJspeert, Pauline A. van +Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, +Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation +of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and +Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_overview.htm --- a/shm_overview.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_overview.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,332 +1,332 @@ - - - - - - - - - - - -
- -

Info -table

- -

This -table contains information on different characteristics of SHM. For all -characteristics information can be found for all sequences or only sequences of -a certain (sub)class. All results are based on the sequences that passed the filter -settings chosen on the start page of the SHM & CSR pipeline and only -include details on the analysed region as determined by the setting of the -sequence starts at filter. All data in this table can be downloaded via the -“downloads” tab.

- -

Mutation -frequency:

- -

These values -give information on the level of SHM. More information -on the values found in healthy individuals of different ages can be found in IJspeert -and van Schouwenburg et al, PMID: 27799928

- -

Number -of mutations: Shows the number of total -mutations / the number of sequenced bases (the % of mutated bases).

- -

Median -number of mutations: Shows the median % of -SHM of all sequences.

- -

Patterns -of SHM:

- -

These values -give insights into the targeting and patterns of SHM. These values can give -insight into the repair pathways used to repair the U:G mismatches introduced -by AID. More information -on the values found in healthy individuals of different ages can be found in -IJspeert and van Schouwenburg et al, PMID: 27799928

- -

Transitions: -Shows the number of transition mutations / the number of total mutations (the -percentage of mutations that are transitions). Transition mutations are C>T, -T>C, A>G, G>A.

- -

Transversions: -Shows the number of transversion mutations / the number of total mutations (the -percentage of mutations that are transitions). Transversion mutations are -C>A, C>G, T>A, T>G, A>T, A>C, G>T, G>C.

- -

Transitions -at GC: Shows the number of transitions at GC locations (C>T, -G>A) / the total number of mutations at GC locations (the percentage of -mutations at GC locations that are transitions).

- -

Targeting -of GC: Shows the number of mutations at GC -locations / the total number of mutations (the percentage of total mutations -that are at GC locations).

- -

Transitions -at AT: Shows the number of transitions at AT -locations (T>C, A>G) / the total number of mutations at AT locations (the -percentage of mutations at AT locations that are transitions).

- -

Targeting -of AT: Shows the number of mutations at AT -locations / the total number of mutations (the percentage of total mutations -that are at AT locations).

- -

RGYW: -Shows -the number of mutations that are in a RGYW motive / The number of total mutations -(the percentage of mutations that are in a RGYW motive). RGYW motives are known to be -preferentially targeted by AID (R=Purine, -Y=pyrimidine, W = A or T).

- -

WRCY: -Shows the number of mutations -that are in a WRCY motive / The number of -total mutations (the percentage of mutations that are in a WRCY motive). WRCY -motives are known to be preferentially targeted by AID (R=Purine, -Y=pyrimidine, W = A or T).

- -

WA: -Shows -the number of mutations that are in a WA motive / The number of total mutations -(the percentage of mutations that are in a WA motive). It is described that -polymerase eta preferentially makes errors at WA motives (W -= A or T).

- -

TW: -Shows the number of mutations that are in a TW motive / The number of total mutations -(the percentage of mutations that are in a TW motive). It is described that -polymerase eta preferentially makes errors at TW motives (W -= A or T).

- -

Antigen -selection:

- -

These -values give insight into antigen selection. It has been described that during -antigen selection, there is selection against replacement mutations in the FR -regions as these can cause instability of the B-cell receptor. In contrast -replacement mutations in the CDR regions are important for changing the -affinity of the B-cell receptor and therefore there is selection for this type -of mutations. Silent mutations do not alter the amino acid sequence and -therefore do not play a role in selection. More information on the values found -in healthy individuals of different ages can be found in IJspeert and van -Schouwenburg et al, PMID: 27799928

- -

FR -R/S: Shows the number of replacement -mutations in the FR regions / The number of silent mutations in the FR regions -(the number of replacement mutations in the FR regions divided by the number of -silent mutations in the FR regions)

- -

CDR -R/S: Shows the number of replacement -mutations in the CDR regions / The number of silent mutations in the CDR -regions (the number of replacement mutations in the CDR regions divided by the -number of silent mutations in the CDR regions)

- -

Number -of sequences nucleotides:

- -

These -values give information on the number of sequenced nucleotides.

- -

Nt -in FR: Shows the number of sequences bases -that are located in the FR regions / The total number of sequenced bases (the -percentage of sequenced bases that are present in the FR regions).

- -

Nt -in CDR: Shows the number of sequenced bases -that are located in the CDR regions / The total number of sequenced bases (the percentage of -sequenced bases that are present in the CDR regions).

- -

A: -Shows the total number of sequenced -adenines / The total number of sequenced bases (the percentage of sequenced -bases that were adenines).

- -

C: -Shows -the total number of sequenced cytosines / The total number of sequenced bases -(the percentage of sequenced bases that were cytosines).

- -

T: -Shows -the total number of sequenced thymines -/ The total number of sequenced bases (the percentage of sequenced bases that -were thymines).

- -

G: -Shows the total number of sequenced guanines / The total number of -sequenced bases (the percentage of sequenced bases that were guanines).

- -

Graphs

- -

These graphs visualize -information on the patterns and targeting of SHM and thereby give information -into the repair pathways used to repair the U:G mismatches introduced by AID. The -data represented in these graphs can be downloaded in the download tab. More -information on the values found in healthy individuals of different ages can be -found in IJspeert and van Schouwenburg et al, PMID: 27799928. -

- -

Percentage -of mutations in AID and pol eta motives

- -

Visualizes -for each -(sub)class the percentage of mutations that are present in AID (RGYW or -WRCY) or polymerase eta motives (WA or TW) in the different subclasses (R=Purine, -Y=pyrimidine, W = A or T).

- -

Relative -mutation patterns

- -

Visualizes -for each (sub)class the distribution of mutations between mutations at AT -locations and transitions or transversions at GC locations.

- -

Absolute -mutation patterns

- -

Visualized -for each (sub)class the percentage of sequenced AT and GC bases that are -mutated. The mutations at GC bases are divided into transition and transversion -mutations.

- -

Hanna IJspeert, Pauline A. van -Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, -Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation -of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and -Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

- -
- - - - + + + + + + + + + + + +
+ +

Info +table

+ +

This +table contains information on different characteristics of SHM. For all +characteristics information can be found for all sequences or only sequences of +a certain (sub)class. All results are based on the sequences that passed the filter +settings chosen on the start page of the SHM & CSR pipeline and only +include details on the analysed region as determined by the setting of the +sequence starts at filter. All data in this table can be downloaded via the +“downloads” tab.

+ +

Mutation +frequency:

+ +

These values +give information on the level of SHM. More information +on the values found in healthy individuals of different ages can be found in IJspeert +and van Schouwenburg et al, PMID: 27799928

+ +

Number +of mutations: Shows the number of total +mutations / the number of sequenced bases (the % of mutated bases).

+ +

Median +number of mutations: Shows the median % of +SHM of all sequences.

+ +

Patterns +of SHM:

+ +

These values +give insights into the targeting and patterns of SHM. These values can give +insight into the repair pathways used to repair the U:G mismatches introduced +by AID. More information +on the values found in healthy individuals of different ages can be found in +IJspeert and van Schouwenburg et al, PMID: 27799928

+ +

Transitions: +Shows the number of transition mutations / the number of total mutations (the +percentage of mutations that are transitions). Transition mutations are C>T, +T>C, A>G, G>A.

+ +

Transversions: +Shows the number of transversion mutations / the number of total mutations (the +percentage of mutations that are transitions). Transversion mutations are +C>A, C>G, T>A, T>G, A>T, A>C, G>T, G>C.

+ +

Transitions +at GC: Shows the number of transitions at GC locations (C>T, +G>A) / the total number of mutations at GC locations (the percentage of +mutations at GC locations that are transitions).

+ +

Targeting +of GC: Shows the number of mutations at GC +locations / the total number of mutations (the percentage of total mutations +that are at GC locations).

+ +

Transitions +at AT: Shows the number of transitions at AT +locations (T>C, A>G) / the total number of mutations at AT locations (the +percentage of mutations at AT locations that are transitions).

+ +

Targeting +of AT: Shows the number of mutations at AT +locations / the total number of mutations (the percentage of total mutations +that are at AT locations).

+ +

RGYW: +Shows +the number of mutations that are in a RGYW motive / The number of total mutations +(the percentage of mutations that are in a RGYW motive). RGYW motives are known to be +preferentially targeted by AID (R=Purine, +Y=pyrimidine, W = A or T).

+ +

WRCY: +Shows the number of mutations +that are in a WRCY motive / The number of +total mutations (the percentage of mutations that are in a WRCY motive). WRCY +motives are known to be preferentially targeted by AID (R=Purine, +Y=pyrimidine, W = A or T).

+ +

WA: +Shows +the number of mutations that are in a WA motive / The number of total mutations +(the percentage of mutations that are in a WA motive). It is described that +polymerase eta preferentially makes errors at WA motives (W += A or T).

+ +

TW: +Shows the number of mutations that are in a TW motive / The number of total mutations +(the percentage of mutations that are in a TW motive). It is described that +polymerase eta preferentially makes errors at TW motives (W += A or T).

+ +

Antigen +selection:

+ +

These +values give insight into antigen selection. It has been described that during +antigen selection, there is selection against replacement mutations in the FR +regions as these can cause instability of the B-cell receptor. In contrast +replacement mutations in the CDR regions are important for changing the +affinity of the B-cell receptor and therefore there is selection for this type +of mutations. Silent mutations do not alter the amino acid sequence and +therefore do not play a role in selection. More information on the values found +in healthy individuals of different ages can be found in IJspeert and van +Schouwenburg et al, PMID: 27799928

+ +

FR +R/S: Shows the number of replacement +mutations in the FR regions / The number of silent mutations in the FR regions +(the number of replacement mutations in the FR regions divided by the number of +silent mutations in the FR regions)

+ +

CDR +R/S: Shows the number of replacement +mutations in the CDR regions / The number of silent mutations in the CDR +regions (the number of replacement mutations in the CDR regions divided by the +number of silent mutations in the CDR regions)

+ +

Number +of sequences nucleotides:

+ +

These +values give information on the number of sequenced nucleotides.

+ +

Nt +in FR: Shows the number of sequences bases +that are located in the FR regions / The total number of sequenced bases (the +percentage of sequenced bases that are present in the FR regions).

+ +

Nt +in CDR: Shows the number of sequenced bases +that are located in the CDR regions / The total number of sequenced bases (the percentage of +sequenced bases that are present in the CDR regions).

+ +

A: +Shows the total number of sequenced +adenines / The total number of sequenced bases (the percentage of sequenced +bases that were adenines).

+ +

C: +Shows +the total number of sequenced cytosines / The total number of sequenced bases +(the percentage of sequenced bases that were cytosines).

+ +

T: +Shows +the total number of sequenced thymines +/ The total number of sequenced bases (the percentage of sequenced bases that +were thymines).

+ +

G: +Shows the total number of sequenced guanines / The total number of +sequenced bases (the percentage of sequenced bases that were guanines).

+ +

Graphs

+ +

These graphs visualize +information on the patterns and targeting of SHM and thereby give information +into the repair pathways used to repair the U:G mismatches introduced by AID. The +data represented in these graphs can be downloaded in the download tab. More +information on the values found in healthy individuals of different ages can be +found in IJspeert and van Schouwenburg et al, PMID: 27799928. +

+ +

Percentage +of mutations in AID and pol eta motives

+ +

Visualizes +for each +(sub)class the percentage of mutations that are present in AID (RGYW or +WRCY) or polymerase eta motives (WA or TW) in the different subclasses (R=Purine, +Y=pyrimidine, W = A or T).

+ +

Relative +mutation patterns

+ +

Visualizes +for each (sub)class the distribution of mutations between mutations at AT +locations and transitions or transversions at GC locations.

+ +

Absolute +mutation patterns

+ +

Visualized +for each (sub)class the percentage of sequenced AT and GC bases that are +mutated. The mutations at GC bases are divided into transition and transversion +mutations.

+ +

Hanna IJspeert, Pauline A. van +Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, +Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation +of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and +Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_selection.htm --- a/shm_selection.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_selection.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,128 +1,128 @@ - - - - - - - - - - - -
- -

References

- -

Yaari, G. and Uduman, M. and Kleinstein, S. H. (2012). Quantifying -selection in high-throughput Immunoglobulin sequencing data sets. In Nucleic Acids Research, 40 (17), -pp. e134–e134. [doi:10.1093/nar/gks457][Link]

- -

Graphs

- -

AA -mutation frequency

- -

For -each class, the frequency of replacement mutations at each amino acid position -is shown, which is calculated by dividing the number of replacement mutations -at a particular amino acid position/the number sequences that have an amino -acid at that particular position. Since the length of the CDR1 and CDR2 region -is not the same for every VH gene, some amino acids positions are absent. -Therefore we calculate the frequency using the number of amino acids present at -that that particular location.

- -

Antigen -selection (BASELINe)

- -

Shows -the results of the analysis of antigen selection as performed using BASELINe. -Details on the analysis performed by BASELINe can be found in Yaari et al, -PMID: 22641856. The settings used for the analysis are: -focused, SHM targeting model: human Tri-nucleotide, custom bounderies. The -custom boundries are dependent on the ‘sequence starts at filter’.

- -

Leader: -1:26:38:55:65:104:-

- -

FR1: 27:27:38:55:65:104:-

- -

CDR1: 27:27:38:55:65:104:-

- -

FR2: 27:27:38:55:65:104:-

- -

Hanna IJspeert, Pauline A. van -Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, -Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation -of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and -Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

- -
- - - - + + + + + + + + + + + +
+ +

References

+ +

Yaari, G. and Uduman, M. and Kleinstein, S. H. (2012). Quantifying +selection in high-throughput Immunoglobulin sequencing data sets. In Nucleic Acids Research, 40 (17), +pp. e134–e134. [doi:10.1093/nar/gks457][Link]

+ +

Graphs

+ +

AA +mutation frequency

+ +

For +each class, the frequency of replacement mutations at each amino acid position +is shown, which is calculated by dividing the number of replacement mutations +at a particular amino acid position/the number sequences that have an amino +acid at that particular position. Since the length of the CDR1 and CDR2 region +is not the same for every VH gene, some amino acids positions are absent. +Therefore we calculate the frequency using the number of amino acids present at +that that particular location.

+ +

Antigen +selection (BASELINe)

+ +

Shows +the results of the analysis of antigen selection as performed using BASELINe. +Details on the analysis performed by BASELINe can be found in Yaari et al, +PMID: 22641856. The settings used for the analysis are: +focused, SHM targeting model: human Tri-nucleotide, custom bounderies. The +custom boundries are dependent on the ‘sequence starts at filter’.

+ +

Leader: +1:26:38:55:65:104:-

+ +

FR1: 27:27:38:55:65:104:-

+ +

CDR1: 27:27:38:55:65:104:-

+ +

FR2: 27:27:38:55:65:104:-

+ +

Hanna IJspeert, Pauline A. van +Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, +Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation +of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and +Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca shm_transition.htm --- a/shm_transition.htm Thu Dec 07 03:44:38 2017 -0500 +++ b/shm_transition.htm Tue Jan 29 03:54:09 2019 -0500 @@ -1,120 +1,120 @@ - - - - - - - - - - - -
- -

These graphs and -tables give insight into the targeting and patterns of SHM. This can give -insight into the DNA repair pathways used to solve the U:G mismatches -introduced by AID. More information on the values found in healthy individuals -of different ages can be found in IJspeert and van Schouwenburg et al, PMID: -27799928.

- -

Graphs -

- -

Heatmap transition -information

- -

Heatmaps visualizing for each subclass the frequency -of all possible substitutions. On the x-axes the original base is shown, while -the y-axes shows the new base. The darker the shade of blue, the more frequent -this type of substitution is occurring. 

- -

Bargraph -transition information

- -

Bar graph -visualizing for each original base the distribution of substitutions into the other -bases. A graph is included for each (sub)class.

- -

Tables

- -

Transition -tables are shown for each (sub)class. All the original bases are listed -horizontally, while the new bases are listed vertically.

- -

Hanna IJspeert, Pauline A. van -Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, -Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation -of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and -Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

- -
- - - - + + + + + + + + + + + +
+ +

These graphs and +tables give insight into the targeting and patterns of SHM. This can give +insight into the DNA repair pathways used to solve the U:G mismatches +introduced by AID. More information on the values found in healthy individuals +of different ages can be found in IJspeert and van Schouwenburg et al, PMID: +27799928.

+ +

Graphs +

+ +

Heatmap transition +information

+ +

Heatmaps visualizing for each subclass the frequency +of all possible substitutions. On the x-axes the original base is shown, while +the y-axes shows the new base. The darker the shade of blue, the more frequent +this type of substitution is occurring. 

+ +

Bargraph +transition information

+ +

Bar graph +visualizing for each original base the distribution of substitutions into the other +bases. A graph is included for each (sub)class.

+ +

Tables

+ +

Transition +tables are shown for each (sub)class. All the original bases are listed +horizontally, while the new bases are listed vertically.

+ +

Hanna IJspeert, Pauline A. van +Schouwenburg, David van Zessen, Ingrid Pico-Knijnenburg, Gertjan J. Driessen, +Andrew P. Stubbs, and Mirjam van der Burg (2016). Evaluation +of the Antigen-Experienced B-Cell Receptor Repertoire in Healthy Children and +Adults. In Frontiers in Immunolog, 7, pp. e410-410. [doi:10.3389/fimmu.2016.00410][Link]

+ +
+ + + + diff -r 43a1aa648537 -r ba33b94637ca wrapper.sh