annotate report_clonality/RScript.r @ 57:33412e85e669 draft

"planemo upload commit 2a9ed3adf321f18047c3746735a5e79a4586798d"
author rhpvorderman
date Wed, 02 Feb 2022 10:50:01 +0000
parents 81b3eb11ed2c
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
54
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1 # ---------------------- load/install packages ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
2
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
3 if (!("gridExtra" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
4 install.packages("gridExtra", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
5 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
6 library(gridExtra)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
7 if (!("ggplot2" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
8 install.packages("ggplot2", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
9 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
10 library(ggplot2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
11 if (!("plyr" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
12 install.packages("plyr", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
13 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
14 library(plyr)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
15
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
16 if (!("data.table" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
17 install.packages("data.table", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
18 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
19 library(data.table)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
20
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
21 if (!("reshape2" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
22 install.packages("reshape2", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
23 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
24 library(reshape2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
25
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
26 if (!("lymphclon" %in% rownames(installed.packages()))) {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
27 install.packages("lymphclon", repos="http://cran.xl-mirror.nl/")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
28 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
29 library(lymphclon)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
30
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
31 # ---------------------- parameters ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
32
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
33 args <- commandArgs(trailingOnly = TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
34
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
35 infile = args[1] #path to input file
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
36 outfile = args[2] #path to output file
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
37 outdir = args[3] #path to output folder (html/images/data)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
38 clonaltype = args[4] #clonaltype definition, or 'none' for no unique filtering
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
39 ct = unlist(strsplit(clonaltype, ","))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
40 species = args[5] #human or mouse
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
41 locus = args[6] # IGH, IGK, IGL, TRB, TRA, TRG or TRD
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
42 filterproductive = ifelse(args[7] == "yes", T, F) #should unproductive sequences be filtered out? (yes/no)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
43 clonality_method = args[8]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
44
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
45
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
46 # ---------------------- Data preperation ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
47
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
48 print("Report Clonality - Data preperation")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
49
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
50 inputdata = read.table(infile, sep="\t", header=TRUE, fill=T, comment.char="", stringsAsFactors=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
51
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
52 inputdata$Sample = as.character(inputdata$Sample)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
53
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
54
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
55 print(paste("nrows: ", nrow(inputdata)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
56
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
57 setwd(outdir)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
58
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
59 # remove weird rows
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
60 inputdata = inputdata[inputdata$Sample != "",]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
61
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
62 print(paste("nrows: ", nrow(inputdata)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
63
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
64 #remove the allele from the V,D and J genes
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
65 inputdata$Top.V.Gene = gsub("[*]([0-9]+)", "", inputdata$Top.V.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
66 inputdata$Top.D.Gene = gsub("[*]([0-9]+)", "", inputdata$Top.D.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
67 inputdata$Top.J.Gene = gsub("[*]([0-9]+)", "", inputdata$Top.J.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
68
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
69 print(paste("nrows: ", nrow(inputdata)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
70
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
71 #filter uniques
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
72 inputdata.removed = inputdata[NULL,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
73
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
74 print(paste("nrows: ", nrow(inputdata)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
75
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
76 inputdata$clonaltype = 1:nrow(inputdata)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
77
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
78 #keep track of the count of sequences in samples or samples/replicates for the front page overview
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
79 input.sample.count = data.frame(data.table(inputdata)[, list(All=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
80 input.rep.count = data.frame(data.table(inputdata)[, list(All=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
81
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
82 PRODF = inputdata
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
83 UNPROD = inputdata
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
84 if(filterproductive){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
85 if("Functionality" %in% colnames(inputdata)) { # "Functionality" is an IMGT column
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
86 #PRODF = inputdata[inputdata$Functionality == "productive" | inputdata$Functionality == "productive (see comment)", ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
87 PRODF = inputdata[inputdata$Functionality %in% c("productive (see comment)","productive"),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
88
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
89 PRODF.count = data.frame(data.table(PRODF)[, list(count=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
90
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
91 UNPROD = inputdata[inputdata$Functionality %in% c("unproductive (see comment)","unproductive"), ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
92 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
93 PRODF = inputdata[inputdata$VDJ.Frame != "In-frame with stop codon" & inputdata$VDJ.Frame != "Out-of-frame" & inputdata$CDR3.Found.How != "NOT_FOUND" , ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
94 UNPROD = inputdata[!(inputdata$VDJ.Frame != "In-frame with stop codon" & inputdata$VDJ.Frame != "Out-of-frame" & inputdata$CDR3.Found.How != "NOT_FOUND" ), ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
95 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
96 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
97
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
98 for(i in 1:nrow(UNPROD)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
99 if(!is.numeric(UNPROD[i,"CDR3.Length"])){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
100 UNPROD[i,"CDR3.Length"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
101 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
102 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
103
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
104 prod.sample.count = data.frame(data.table(PRODF)[, list(Productive=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
105 prod.rep.count = data.frame(data.table(PRODF)[, list(Productive=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
106
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
107 unprod.sample.count = data.frame(data.table(UNPROD)[, list(Unproductive=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
108 unprod.rep.count = data.frame(data.table(UNPROD)[, list(Unproductive=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
109
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
110 clonalityFrame = PRODF
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
111
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
112 #remove duplicates based on the clonaltype
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
113 if(clonaltype != "none"){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
114 clonaltype = paste(clonaltype, ",Sample", sep="") #add sample column to clonaltype, unique within samples
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
115 PRODF$clonaltype = do.call(paste, c(PRODF[unlist(strsplit(clonaltype, ","))], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
116 PRODF = PRODF[!duplicated(PRODF$clonaltype), ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
117
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
118 UNPROD$clonaltype = do.call(paste, c(UNPROD[unlist(strsplit(clonaltype, ","))], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
119 UNPROD = UNPROD[!duplicated(UNPROD$clonaltype), ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
120
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
121 #again for clonalityFrame but with sample+replicate
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
122 clonalityFrame$clonaltype = do.call(paste, c(clonalityFrame[unlist(strsplit(clonaltype, ","))], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
123 clonalityFrame$clonality_clonaltype = do.call(paste, c(clonalityFrame[unlist(strsplit(paste(clonaltype, ",Replicate", sep=""), ","))], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
124 clonalityFrame = clonalityFrame[!duplicated(clonalityFrame$clonality_clonaltype), ]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
125 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
126
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
127 if(nrow(PRODF) == 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
128 stop("No sequences left after filtering")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
129 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
130
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
131 prod.unique.sample.count = data.frame(data.table(PRODF)[, list(Productive_unique=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
132 prod.unique.rep.count = data.frame(data.table(PRODF)[, list(Productive_unique=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
133
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
134 unprod.unique.sample.count = data.frame(data.table(UNPROD)[, list(Unproductive_unique=.N), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
135 unprod.unique.rep.count = data.frame(data.table(UNPROD)[, list(Unproductive_unique=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
136
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
137
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
138 PRODF$freq = 1
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
139
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
140 if(any(grepl(pattern="_", x=PRODF$ID))){ #the frequency can be stored in the ID with the pattern ".*_freq_.*"
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
141 PRODF$freq = gsub("^[0-9]+_", "", PRODF$ID)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
142 PRODF$freq = gsub("_.*", "", PRODF$freq)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
143 PRODF$freq = as.numeric(PRODF$freq)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
144 if(any(is.na(PRODF$freq))){ #if there was an "_" in the ID, but not the frequency, go back to frequency of 1 for every sequence
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
145 PRODF$freq = 1
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
146 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
147 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
148
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
149 #make a names list with sample -> color
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
150 naive.colors = c('blue4', 'darkred', 'olivedrab3', 'red', 'gray74', 'darkviolet', 'lightblue1', 'gold', 'chartreuse2', 'pink', 'Paleturquoise3', 'Chocolate1', 'Yellow', 'Deeppink3', 'Mediumorchid1', 'Darkgreen', 'Blue', 'Gray36', 'Hotpink', 'Yellow4')
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
151 unique.samples = unique(PRODF$Sample)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
152
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
153 if(length(unique.samples) <= length(naive.colors)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
154 sample.colors = naive.colors[1:length(unique.samples)]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
155 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
156 sample.colors = rainbow(length(unique.samples))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
157 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
158
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
159 names(sample.colors) = unique.samples
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
160
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
161 print("Sample.colors")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
162 print(sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
163
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
164
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
165 #write the complete dataset that is left over, will be the input if 'none' for clonaltype and 'no' for filterproductive
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
166 write.table(PRODF, "allUnique.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
167 #write.table(PRODF, "allUnique.csv", sep=",",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
168 write.table(UNPROD, "allUnproductive.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
169
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
170 print("SAMPLE TABLE:")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
171 print(table(PRODF$Sample))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
172
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
173 #write the samples to a file
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
174 sampleFile <- file("samples.txt")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
175 un = unique(inputdata$Sample)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
176 un = paste(un, sep="\n")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
177 writeLines(un, sampleFile)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
178 close(sampleFile)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
179
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
180 # ---------------------- Counting the productive/unproductive and unique sequences ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
181
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
182 print("Report Clonality - counting productive/unproductive/unique")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
183
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
184 #create the table on the overview page with the productive/unique counts per sample/replicate
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
185 #first for sample
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
186
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
187 sample.count = merge(input.sample.count, prod.sample.count, by="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
188 sample.count$perc_prod = round(sample.count$Productive / sample.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
189 sample.count = merge(sample.count, prod.unique.sample.count, by="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
190 sample.count$perc_prod_un = round(sample.count$Productive_unique / sample.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
191
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
192 sample.count = merge(sample.count , unprod.sample.count, by="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
193 sample.count$perc_unprod = round(sample.count$Unproductive / sample.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
194 sample.count = merge(sample.count, unprod.unique.sample.count, by="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
195 sample.count$perc_unprod_un = round(sample.count$Unproductive_unique / sample.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
196
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
197 #then sample/replicate
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
198 rep.count = merge(input.rep.count, prod.rep.count, by=c("Sample", "Replicate"), all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
199
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
200 print(rep.count)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
201
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
202 fltr = is.na(rep.count$Productive)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
203 if(any(fltr)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
204 rep.count[fltr,"Productive"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
205 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
206
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
207 print(rep.count)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
208
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
209 rep.count$perc_prod = round(rep.count$Productive / rep.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
210 rep.count = merge(rep.count, prod.unique.rep.count, by=c("Sample", "Replicate"), all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
211 rep.count$perc_prod_un = round(rep.count$Productive_unique / rep.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
212
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
213 rep.count = merge(rep.count, unprod.rep.count, by=c("Sample", "Replicate"), all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
214 rep.count$perc_unprod = round(rep.count$Unproductive / rep.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
215 rep.count = merge(rep.count, unprod.unique.rep.count, by=c("Sample", "Replicate"), all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
216 rep.count$perc_unprod_un = round(rep.count$Unproductive_unique / rep.count$All * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
217
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
218 rep.count$Sample = paste(rep.count$Sample, rep.count$Replicate, sep="_")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
219 rep.count = rep.count[,names(rep.count) != "Replicate"]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
220
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
221 count = rbind(sample.count, rep.count)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
222
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
223
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
224
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
225 write.table(x=count, file="productive_counting.txt", sep=",",quote=F,row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
226
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
227 # ---------------------- V+J+CDR3 sequence count ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
228
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
229 VJCDR3.count = data.frame(table(clonalityFrame$Top.V.Gene, clonalityFrame$Top.J.Gene, clonalityFrame$CDR3.Seq.DNA))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
230 names(VJCDR3.count) = c("Top.V.Gene", "Top.J.Gene", "CDR3.Seq.DNA", "Count")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
231
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
232 VJCDR3.count = VJCDR3.count[VJCDR3.count$Count > 0,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
233 VJCDR3.count = VJCDR3.count[order(-VJCDR3.count$Count),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
234
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
235 write.table(x=VJCDR3.count, file="VJCDR3_count.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
236
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
237 # ---------------------- Frequency calculation for V, D and J ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
238
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
239 print("Report Clonality - frequency calculation V, D and J")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
240
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
241 PRODFV = data.frame(data.table(PRODF)[, list(Length=sum(freq)), by=c("Sample", "Top.V.Gene")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
242 Total = ddply(PRODFV, .(Sample), function(x) data.frame(Total = sum(x$Length)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
243 PRODFV = merge(PRODFV, Total, by.x='Sample', by.y='Sample', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
244 PRODFV = ddply(PRODFV, c("Sample", "Top.V.Gene"), summarise, relFreq= (Length*100 / Total))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
245
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
246 PRODFD = data.frame(data.table(PRODF)[, list(Length=sum(freq)), by=c("Sample", "Top.D.Gene")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
247 Total = ddply(PRODFD, .(Sample), function(x) data.frame(Total = sum(x$Length)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
248 PRODFD = merge(PRODFD, Total, by.x='Sample', by.y='Sample', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
249 PRODFD = ddply(PRODFD, c("Sample", "Top.D.Gene"), summarise, relFreq= (Length*100 / Total))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
250
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
251 PRODFJ = data.frame(data.table(PRODF)[, list(Length=sum(freq)), by=c("Sample", "Top.J.Gene")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
252 Total = ddply(PRODFJ, .(Sample), function(x) data.frame(Total = sum(x$Length)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
253 PRODFJ = merge(PRODFJ, Total, by.x='Sample', by.y='Sample', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
254 PRODFJ = ddply(PRODFJ, c("Sample", "Top.J.Gene"), summarise, relFreq= (Length*100 / Total))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
255
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
256 # ---------------------- Setting up the gene names for the different species/loci ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
257
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
258 print("Report Clonality - getting genes for species/loci")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
259
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
260 Vchain = ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
261 Dchain = ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
262 Jchain = ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
263
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
264 if(species == "custom"){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
265 print("Custom genes: ")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
266 splt = unlist(strsplit(locus, ";"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
267 print(paste("V:", splt[1]))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
268 print(paste("D:", splt[2]))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
269 print(paste("J:", splt[3]))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
270
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
271 Vchain = unlist(strsplit(splt[1], ","))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
272 Vchain = data.frame(v.name = Vchain, chr.orderV = 1:length(Vchain))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
273
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
274 Dchain = unlist(strsplit(splt[2], ","))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
275 if(length(Dchain) > 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
276 Dchain = data.frame(v.name = Dchain, chr.orderD = 1:length(Dchain))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
277 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
278 Dchain = data.frame(v.name = character(0), chr.orderD = numeric(0))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
279 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
280
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
281 Jchain = unlist(strsplit(splt[3], ","))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
282 Jchain = data.frame(v.name = Jchain, chr.orderJ = 1:length(Jchain))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
283
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
284 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
285 genes = read.table("genes.txt", sep="\t", header=TRUE, fill=T, comment.char="")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
286
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
287 Vchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "V",c("IMGT.GENE.DB", "chr.order")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
288 colnames(Vchain) = c("v.name", "chr.orderV")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
289 Dchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "D",c("IMGT.GENE.DB", "chr.order")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
290 colnames(Dchain) = c("v.name", "chr.orderD")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
291 Jchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "J",c("IMGT.GENE.DB", "chr.order")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
292 colnames(Jchain) = c("v.name", "chr.orderJ")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
293 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
294 useD = TRUE
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
295 if(nrow(Dchain) == 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
296 useD = FALSE
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
297 cat("No D Genes in this species/locus")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
298 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
299 print(paste(nrow(Vchain), "genes in V"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
300 print(paste(nrow(Dchain), "genes in D"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
301 print(paste(nrow(Jchain), "genes in J"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
302
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
303 # ---------------------- merge with the frequency count ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
304
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
305 PRODFV = merge(PRODFV, Vchain, by.x='Top.V.Gene', by.y='v.name', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
306
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
307 PRODFD = merge(PRODFD, Dchain, by.x='Top.D.Gene', by.y='v.name', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
308
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
309 PRODFJ = merge(PRODFJ, Jchain, by.x='Top.J.Gene', by.y='v.name', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
310
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
311 # ---------------------- Create the V, D and J frequency plots and write the data.frame for every plot to a file ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
312
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
313 print("Report Clonality - V, D and J frequency plots")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
314
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
315 pV = ggplot(PRODFV)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
316 pV = pV + geom_bar( aes( x=factor(reorder(Top.V.Gene, chr.orderV)), y=relFreq, fill=Sample), stat='identity', position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
317 pV = pV + xlab("Summary of V gene") + ylab("Frequency") + ggtitle("Relative frequency of V gene usage") + scale_fill_manual(values=sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
318 pV = pV + theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
319 write.table(x=PRODFV, file="VFrequency.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
320
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
321 png("VPlot.png",width = 1280, height = 720)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
322 pV
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
323 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
324
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
325 ggsave("VPlot.pdf", pV, width=13, height=7)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
326
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
327 if(useD){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
328 pD = ggplot(PRODFD)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
329 pD = pD + geom_bar( aes( x=factor(reorder(Top.D.Gene, chr.orderD)), y=relFreq, fill=Sample), stat='identity', position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
330 pD = pD + xlab("Summary of D gene") + ylab("Frequency") + ggtitle("Relative frequency of D gene usage") + scale_fill_manual(values=sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
331 pD = pD + theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
332 write.table(x=PRODFD, file="DFrequency.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
333
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
334 png("DPlot.png",width = 800, height = 600)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
335 print(pD)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
336 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
337
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
338 ggsave("DPlot.pdf", pD, width=10, height=7)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
339 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
340
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
341 pJ = ggplot(PRODFJ)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
342 pJ = pJ + geom_bar( aes( x=factor(reorder(Top.J.Gene, chr.orderJ)), y=relFreq, fill=Sample), stat='identity', position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
343 pJ = pJ + xlab("Summary of J gene") + ylab("Frequency") + ggtitle("Relative frequency of J gene usage") + scale_fill_manual(values=sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
344 pJ = pJ + theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
345 write.table(x=PRODFJ, file="JFrequency.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
346
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
347 png("JPlot.png",width = 800, height = 600)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
348 pJ
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
349 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
350
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
351 ggsave("JPlot.pdf", pJ)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
352
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
353 # ---------------------- Now the frequency plots of the V, D and J families ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
354
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
355 print("Report Clonality - V, D and J family plots")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
356
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
357 VGenes = PRODF[,c("Sample", "Top.V.Gene")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
358 VGenes$Top.V.Gene = gsub("-.*", "", VGenes$Top.V.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
359 VGenes = data.frame(data.table(VGenes)[, list(Count=.N), by=c("Sample", "Top.V.Gene")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
360 TotalPerSample = data.frame(data.table(VGenes)[, list(total=sum(.SD$Count)), by=Sample])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
361 VGenes = merge(VGenes, TotalPerSample, by="Sample")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
362 VGenes$Frequency = VGenes$Count * 100 / VGenes$total
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
363 VPlot = ggplot(VGenes)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
364 VPlot = VPlot + geom_bar(aes( x = Top.V.Gene, y = Frequency, fill = Sample), stat='identity', position='dodge' ) + theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
365 ggtitle("Distribution of V gene families") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
366 ylab("Percentage of sequences") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
367 scale_fill_manual(values=sample.colors) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
368 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
369 png("VFPlot.png")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
370 VPlot
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
371 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
372 ggsave("VFPlot.pdf", VPlot)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
373
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
374 write.table(x=VGenes, file="VFFrequency.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
375
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
376 if(useD){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
377 DGenes = PRODF[,c("Sample", "Top.D.Gene")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
378 DGenes$Top.D.Gene = gsub("-.*", "", DGenes$Top.D.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
379 DGenes = data.frame(data.table(DGenes)[, list(Count=.N), by=c("Sample", "Top.D.Gene")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
380 TotalPerSample = data.frame(data.table(DGenes)[, list(total=sum(.SD$Count)), by=Sample])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
381 DGenes = merge(DGenes, TotalPerSample, by="Sample")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
382 DGenes$Frequency = DGenes$Count * 100 / DGenes$total
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
383 DPlot = ggplot(DGenes)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
384 DPlot = DPlot + geom_bar(aes( x = Top.D.Gene, y = Frequency, fill = Sample), stat='identity', position='dodge' ) + theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
385 ggtitle("Distribution of D gene families") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
386 ylab("Percentage of sequences") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
387 scale_fill_manual(values=sample.colors) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
388 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
389 png("DFPlot.png")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
390 print(DPlot)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
391 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
392
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
393 ggsave("DFPlot.pdf", DPlot)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
394 write.table(x=DGenes, file="DFFrequency.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
395 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
396
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
397 # ---------------------- Plotting the cdr3 length ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
398
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
399 print("Report Clonality - CDR3 length plot")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
400
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
401 CDR3Length = data.frame(data.table(PRODF)[, list(Count=.N), by=c("Sample", "CDR3.Length")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
402 TotalPerSample = data.frame(data.table(CDR3Length)[, list(total=sum(.SD$Count)), by=Sample])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
403 CDR3Length = merge(CDR3Length, TotalPerSample, by="Sample")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
404 CDR3Length$Frequency = CDR3Length$Count * 100 / CDR3Length$total
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
405 CDR3LengthPlot = ggplot(CDR3Length)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
406 CDR3LengthPlot = CDR3LengthPlot + geom_bar(aes( x = factor(reorder(CDR3.Length, as.numeric(CDR3.Length))), y = Frequency, fill = Sample), stat='identity', position='dodge' ) + theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
407 ggtitle("Length distribution of CDR3") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
408 xlab("CDR3 Length") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
409 ylab("Percentage of sequences") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
410 scale_fill_manual(values=sample.colors) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
411 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
412 png("CDR3LengthPlot.png",width = 1280, height = 720)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
413 CDR3LengthPlot
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
414 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
415
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
416 ggsave("CDR3LengthPlot.pdf", CDR3LengthPlot, width=12, height=7)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
417
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
418 write.table(x=CDR3Length, file="CDR3LengthPlot.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
419
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
420 # ---------------------- Plot the heatmaps ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
421
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
422 #get the reverse order for the V and D genes
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
423 revVchain = Vchain
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
424 revDchain = Dchain
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
425 revVchain$chr.orderV = rev(revVchain$chr.orderV)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
426 revDchain$chr.orderD = rev(revDchain$chr.orderD)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
427
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
428 if(useD){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
429 print("Report Clonality - Heatmaps VD")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
430 plotVD <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
431 if(length(dat[,1]) == 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
432 return()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
433 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
434
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
435 img = ggplot() +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
436 geom_tile(data=dat, aes(x=factor(reorder(Top.D.Gene, chr.orderD)), y=factor(reorder(Top.V.Gene, chr.orderV)), fill=relLength)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
437 theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
438 scale_fill_gradient(low="gold", high="blue", na.value="white") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
439 ggtitle(paste(unique(dat$Sample), " (N=" , sum(dat$Length, na.rm=T) ,")", sep="")) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
440 xlab("D genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
441 ylab("V Genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
442 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), panel.grid.major = element_line(colour = "gainsboro"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
443
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
444 png(paste("HeatmapVD_", unique(dat[3])[1,1] , ".png", sep=""), width=200+(15*length(Dchain$v.name)), height=100+(15*length(Vchain$v.name)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
445 print(img)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
446 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
447
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
448 ggsave(paste("HeatmapVD_", unique(dat[3])[1,1] , ".pdf", sep=""), img, height=13, width=8)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
449
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
450 write.table(x=acast(dat, Top.V.Gene~Top.D.Gene, value.var="Length"), file=paste("HeatmapVD_", unique(dat[3])[1,1], ".txt", sep=""), sep="\t",quote=F,row.names=T,col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
451 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
452
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
453 VandDCount = data.frame(data.table(PRODF)[, list(Length=.N), by=c("Top.V.Gene", "Top.D.Gene", "Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
454
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
455 VandDCount$l = log(VandDCount$Length)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
456 maxVD = data.frame(data.table(VandDCount)[, list(max=max(l)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
457 VandDCount = merge(VandDCount, maxVD, by.x="Sample", by.y="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
458 VandDCount$relLength = VandDCount$l / VandDCount$max
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
459 check = is.nan(VandDCount$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
460 if(any(check)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
461 VandDCount[check,"relLength"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
462 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
463
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
464 completeVD = merge(VandDCount, revVchain, by.x="Top.V.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
465 completeVD = merge(completeVD, Dchain, by.x="Top.D.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
466
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
467 fltr = is.nan(completeVD$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
468 if(all(fltr)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
469 completeVD[fltr,"relLength"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
470 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
471
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
472 VDList = split(completeVD, f=completeVD[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
473 lapply(VDList, FUN=plotVD)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
474 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
475
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
476 print("Report Clonality - Heatmaps VJ")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
477
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
478 plotVJ <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
479 if(length(dat[,1]) == 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
480 return()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
481 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
482 cat(paste(unique(dat[3])[1,1]))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
483 img = ggplot() +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
484 geom_tile(data=dat, aes(x=factor(reorder(Top.J.Gene, chr.orderJ)), y=factor(reorder(Top.V.Gene, chr.orderV)), fill=relLength)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
485 theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
486 scale_fill_gradient(low="gold", high="blue", na.value="white") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
487 ggtitle(paste(unique(dat$Sample), " (N=" , sum(dat$Length, na.rm=T) ,")", sep="")) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
488 xlab("J genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
489 ylab("V Genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
490 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), panel.grid.major = element_line(colour = "gainsboro"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
491
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
492 png(paste("HeatmapVJ_", unique(dat[3])[1,1] , ".png", sep=""), width=200+(15*length(Jchain$v.name)), height=100+(15*length(Vchain$v.name)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
493 print(img)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
494 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
495
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
496 ggsave(paste("HeatmapVJ_", unique(dat[3])[1,1] , ".pdf", sep=""), img, height=11, width=4)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
497
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
498 write.table(x=acast(dat, Top.V.Gene~Top.J.Gene, value.var="Length"), file=paste("HeatmapVJ_", unique(dat[3])[1,1], ".txt", sep=""), sep="\t",quote=F,row.names=T,col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
499 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
500
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
501
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
502
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
503 VandJCount = data.frame(data.table(PRODF)[, list(Length=.N), by=c("Top.V.Gene", "Top.J.Gene", "Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
504
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
505 VandJCount$l = log(VandJCount$Length)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
506 maxVJ = data.frame(data.table(VandJCount)[, list(max=max(l)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
507 VandJCount = merge(VandJCount, maxVJ, by.x="Sample", by.y="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
508 VandJCount$relLength = VandJCount$l / VandJCount$max
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
509
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
510 check = is.nan(VandJCount$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
511 if(any(check)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
512 VandJCount[check,"relLength"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
513 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
514
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
515 completeVJ = merge(VandJCount, revVchain, by.x="Top.V.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
516 completeVJ = merge(completeVJ, Jchain, by.x="Top.J.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
517
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
518 fltr = is.nan(completeVJ$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
519 if(any(fltr)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
520 completeVJ[fltr,"relLength"] = 1
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
521 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
522
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
523 VJList = split(completeVJ, f=completeVJ[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
524 lapply(VJList, FUN=plotVJ)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
525
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
526
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
527
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
528 if(useD){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
529 print("Report Clonality - Heatmaps DJ")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
530 plotDJ <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
531 if(length(dat[,1]) == 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
532 return()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
533 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
534 img = ggplot() +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
535 geom_tile(data=dat, aes(x=factor(reorder(Top.J.Gene, chr.orderJ)), y=factor(reorder(Top.D.Gene, chr.orderD)), fill=relLength)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
536 theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
537 scale_fill_gradient(low="gold", high="blue", na.value="white") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
538 ggtitle(paste(unique(dat$Sample), " (N=" , sum(dat$Length, na.rm=T) ,")", sep="")) +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
539 xlab("J genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
540 ylab("D Genes") +
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
541 theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), panel.grid.major = element_line(colour = "gainsboro"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
542
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
543 png(paste("HeatmapDJ_", unique(dat[3])[1,1] , ".png", sep=""), width=200+(15*length(Jchain$v.name)), height=100+(15*length(Dchain$v.name)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
544 print(img)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
545 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
546
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
547 ggsave(paste("HeatmapDJ_", unique(dat[3])[1,1] , ".pdf", sep=""), img, width=4, height=7)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
548
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
549 write.table(x=acast(dat, Top.D.Gene~Top.J.Gene, value.var="Length"), file=paste("HeatmapDJ_", unique(dat[3])[1,1], ".txt", sep=""), sep="\t",quote=F,row.names=T,col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
550 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
551
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
552
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
553 DandJCount = data.frame(data.table(PRODF)[, list(Length=.N), by=c("Top.D.Gene", "Top.J.Gene", "Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
554
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
555 DandJCount$l = log(DandJCount$Length)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
556 maxDJ = data.frame(data.table(DandJCount)[, list(max=max(l)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
557 DandJCount = merge(DandJCount, maxDJ, by.x="Sample", by.y="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
558 DandJCount$relLength = DandJCount$l / DandJCount$max
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
559
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
560 check = is.nan(DandJCount$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
561 if(any(check)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
562 DandJCount[check,"relLength"] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
563 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
564
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
565 cartegianProductDJ = expand.grid(Top.D.Gene = Dchain$v.name, Top.J.Gene = Jchain$v.name)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
566
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
567 completeDJ = merge(DandJCount, revDchain, by.x="Top.D.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
568 completeDJ = merge(completeDJ, Jchain, by.x="Top.J.Gene", by.y="v.name", all=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
569
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
570 fltr = is.nan(completeDJ$relLength)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
571 if(any(fltr)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
572 completeDJ[fltr, "relLength"] = 1
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
573 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
574
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
575 DJList = split(completeDJ, f=completeDJ[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
576 lapply(DJList, FUN=plotDJ)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
577 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
578
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
579
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
580 # ---------------------- output tables for the circos plots ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
581
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
582 print("Report Clonality - Circos data")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
583
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
584 for(smpl in unique(PRODF$Sample)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
585 PRODF.sample = PRODF[PRODF$Sample == smpl,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
586
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
587 fltr = PRODF.sample$Top.V.Gene == ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
588 if(any(fltr, na.rm=T)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
589 PRODF.sample[fltr, "Top.V.Gene"] = "NA"
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
590 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
591
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
592 fltr = PRODF.sample$Top.D.Gene == ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
593 if(any(fltr, na.rm=T)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
594 PRODF.sample[fltr, "Top.D.Gene"] = "NA"
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
595 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
596
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
597 fltr = PRODF.sample$Top.J.Gene == ""
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
598 if(any(fltr, na.rm=T)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
599 PRODF.sample[fltr, "Top.J.Gene"] = "NA"
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
600 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
601
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
602 v.d = table(PRODF.sample$Top.V.Gene, PRODF.sample$Top.D.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
603 v.j = table(PRODF.sample$Top.V.Gene, PRODF.sample$Top.J.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
604 d.j = table(PRODF.sample$Top.D.Gene, PRODF.sample$Top.J.Gene)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
605
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
606 write.table(v.d, file=paste(smpl, "_VD_circos.txt", sep=""), sep="\t", quote=F, row.names=T, col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
607 write.table(v.j, file=paste(smpl, "_VJ_circos.txt", sep=""), sep="\t", quote=F, row.names=T, col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
608 write.table(d.j, file=paste(smpl, "_DJ_circos.txt", sep=""), sep="\t", quote=F, row.names=T, col.names=NA)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
609 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
610
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
611 # ---------------------- calculating the clonality score ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
612
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
613 if("Replicate" %in% colnames(inputdata)) #can only calculate clonality score when replicate information is available
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
614 {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
615 print("Report Clonality - Clonality")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
616 write.table(clonalityFrame, "clonalityComplete.txt", sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
617 if(clonality_method == "boyd"){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
618 samples = split(clonalityFrame, clonalityFrame$Sample, drop=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
619
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
620 for (sample in samples){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
621 res = data.frame(paste=character(0))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
622 sample_id = unique(sample$Sample)[[1]]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
623 for(replicate in unique(sample$Replicate)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
624 tmp = sample[sample$Replicate == replicate,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
625 clone_table = data.frame(table(tmp$clonaltype))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
626 clone_col_name = paste("V", replicate, sep="")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
627 colnames(clone_table) = c("paste", clone_col_name)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
628 res = merge(res, clone_table, by="paste", all=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
629 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
630
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
631 res[is.na(res)] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
632
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
633 write.table(res, file=paste("raw_clonality_", sample_id, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
634 write.table(as.matrix(res[,2:ncol(res)]), file=paste("raw_clonality2_", sample_id, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
635
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
636 res = read.table(paste("raw_clonality_", sample_id, ".txt", sep=""), header=F, sep="\t", quote="", stringsAsFactors=F, fill=T, comment.char="")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
637
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
638 infer.result = infer.clonality(as.matrix(res[,2:ncol(res)]))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
639
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
640 #print(infer.result)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
641
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
642 write.table(data.table(infer.result[[12]]), file=paste("lymphclon_clonality_", sample_id, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
643
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
644 res$type = rowSums(res[,2:ncol(res)])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
645
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
646 coincidence.table = data.frame(table(res$type))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
647 colnames(coincidence.table) = c("Coincidence Type", "Raw Coincidence Freq")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
648 write.table(coincidence.table, file=paste("lymphclon_coincidences_", sample_id, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
649 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
650 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
651 clonalFreq = data.frame(data.table(clonalityFrame)[, list(Type=.N), by=c("Sample", "clonaltype")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
652
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
653 #write files for every coincidence group of >1
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
654 samples = unique(clonalFreq$Sample)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
655 for(sample in samples){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
656 clonalFreqSample = clonalFreq[clonalFreq$Sample == sample,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
657 if(max(clonalFreqSample$Type) > 1){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
658 for(i in 2:max(clonalFreqSample$Type)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
659 clonalFreqSampleType = clonalFreqSample[clonalFreqSample$Type == i,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
660 clonalityFrame.sub = clonalityFrame[clonalityFrame$clonaltype %in% clonalFreqSampleType$clonaltype,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
661 clonalityFrame.sub = clonalityFrame.sub[order(clonalityFrame.sub$clonaltype),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
662 write.table(clonalityFrame.sub, file=paste("coincidences_", sample, "_", i, ".txt", sep=""), sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
663 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
664 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
665 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
666
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
667 clonalFreqCount = data.frame(data.table(clonalFreq)[, list(Count=.N), by=c("Sample", "Type")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
668 clonalFreqCount$realCount = clonalFreqCount$Type * clonalFreqCount$Count
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
669 clonalSum = data.frame(data.table(clonalFreqCount)[, list(Reads=sum(realCount)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
670 clonalFreqCount = merge(clonalFreqCount, clonalSum, by.x="Sample", by.y="Sample")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
671
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
672 ct = c('Type\tWeight\n2\t1\n3\t3\n4\t6\n5\t10\n6\t15')
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
673 tcct = textConnection(ct)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
674 CT = read.table(tcct, sep="\t", header=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
675 close(tcct)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
676 clonalFreqCount = merge(clonalFreqCount, CT, by.x="Type", by.y="Type", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
677 clonalFreqCount$WeightedCount = clonalFreqCount$Count * clonalFreqCount$Weight
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
678
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
679 ReplicateReads = data.frame(data.table(clonalityFrame)[, list(Type=.N), by=c("Sample", "Replicate", "clonaltype")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
680 ReplicateReads = data.frame(data.table(ReplicateReads)[, list(Reads=.N), by=c("Sample", "Replicate")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
681 clonalFreqCount$Reads = as.numeric(clonalFreqCount$Reads)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
682 ReplicateReads$Reads = as.numeric(ReplicateReads$Reads)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
683 ReplicateReads$squared = as.numeric(ReplicateReads$Reads * ReplicateReads$Reads)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
684
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
685 ReplicatePrint <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
686 write.table(dat[-1], paste("ReplicateReads_", unique(dat[1])[1,1] , ".txt", sep=""), sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
687 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
688
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
689 ReplicateSplit = split(ReplicateReads, f=ReplicateReads[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
690 lapply(ReplicateSplit, FUN=ReplicatePrint)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
691
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
692 ReplicateReads = data.frame(data.table(ReplicateReads)[, list(ReadsSum=sum(as.numeric(Reads)), ReadsSquaredSum=sum(as.numeric(squared))), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
693 clonalFreqCount = merge(clonalFreqCount, ReplicateReads, by.x="Sample", by.y="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
694
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
695 ReplicateSumPrint <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
696 write.table(dat[-1], paste("ReplicateSumReads_", unique(dat[1])[1,1] , ".txt", sep=""), sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
697 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
698
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
699 ReplicateSumSplit = split(ReplicateReads, f=ReplicateReads[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
700 lapply(ReplicateSumSplit, FUN=ReplicateSumPrint)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
701
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
702 clonalFreqCountSum = data.frame(data.table(clonalFreqCount)[, list(Numerator=sum(WeightedCount, na.rm=T)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
703 clonalFreqCount = merge(clonalFreqCount, clonalFreqCountSum, by.x="Sample", by.y="Sample", all.x=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
704 clonalFreqCount$ReadsSum = as.numeric(clonalFreqCount$ReadsSum) #prevent integer overflow
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
705 clonalFreqCount$Denominator = (((clonalFreqCount$ReadsSum * clonalFreqCount$ReadsSum) - clonalFreqCount$ReadsSquaredSum) / 2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
706 clonalFreqCount$Result = (clonalFreqCount$Numerator + 1) / (clonalFreqCount$Denominator + 1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
707
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
708 ClonalityScorePrint <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
709 write.table(dat$Result, paste("ClonalityScore_", unique(dat[1])[1,1] , ".txt", sep=""), sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
710 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
711
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
712 clonalityScore = clonalFreqCount[c("Sample", "Result")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
713 clonalityScore = unique(clonalityScore)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
714
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
715 clonalityScoreSplit = split(clonalityScore, f=clonalityScore[,"Sample"])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
716 lapply(clonalityScoreSplit, FUN=ClonalityScorePrint)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
717
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
718 clonalityOverview = clonalFreqCount[c("Sample", "Type", "Count", "Weight", "WeightedCount")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
719
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
720
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
721
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
722 ClonalityOverviewPrint <- function(dat){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
723 dat = dat[order(dat[,2]),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
724 write.table(dat[-1], paste("ClonalityOverView_", unique(dat[1])[1,1] , ".txt", sep=""), sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
725 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
726
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
727 clonalityOverviewSplit = split(clonalityOverview, f=clonalityOverview$Sample)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
728 lapply(clonalityOverviewSplit, FUN=ClonalityOverviewPrint)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
729
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
730 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
731
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
732 bak = PRODF
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
733 bakun = UNPROD
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
734
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
735 imgtcolumns = c("X3V.REGION.trimmed.nt.nb","P3V.nt.nb", "N1.REGION.nt.nb", "P5D.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "P3D.nt.nb", "N2.REGION.nt.nb", "P5J.nt.nb", "X5J.REGION.trimmed.nt.nb", "X3V.REGION.trimmed.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "P3V.nt.nb", "P5D.nt.nb", "P3D.nt.nb", "P5J.nt.nb")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
736 if(all(imgtcolumns %in% colnames(inputdata)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
737 {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
738 print("found IMGT columns, running junction analysis")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
739
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
740 #ensure certain columns are in the data (files generated with older versions of IMGT Loader)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
741 col.checks = c("N.REGION.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
742 for(col.check in col.checks){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
743 if(!(col.check %in% names(PRODF))){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
744 print(paste(col.check, "not found adding new column"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
745 if(nrow(PRODF) > 0){ #because R is anoying...
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
746 PRODF[,col.check] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
747 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
748 PRODF = cbind(PRODF, data.frame(N3.REGION.nt.nb=numeric(0), N4.REGION.nt.nb=numeric(0)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
749 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
750 if(nrow(UNPROD) > 0){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
751 UNPROD[,col.check] = 0
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
752 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
753 UNPROD = cbind(UNPROD, data.frame(N3.REGION.nt.nb=numeric(0), N4.REGION.nt.nb=numeric(0)))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
754 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
755 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
756 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
757
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
758 PRODF.with.D = PRODF[nchar(PRODF$Top.D.Gene, keepNA=F) > 2,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
759 PRODF.no.D = PRODF[nchar(PRODF$Top.D.Gene, keepNA=F) < 4,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
760 write.table(PRODF.no.D, "productive_no_D.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
761
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
762 UNPROD.with.D = UNPROD[nchar(UNPROD$Top.D.Gene, keepNA=F) > 2,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
763 UNPROD.no.D = UNPROD[nchar(UNPROD$Top.D.Gene, keepNA=F) < 4,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
764 write.table(UNPROD.no.D, "unproductive_no_D.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
765
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
766 num_median = function(x, na.rm=T) { as.numeric(median(x, na.rm=na.rm)) }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
767
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
768 newData = data.frame(data.table(PRODF.with.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
769 VH.DEL=mean(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
770 P1=mean(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
771 N1=mean(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
772 P2=mean(.SD$P5D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
773 DEL.DH=mean(.SD$X5D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
774 DH.DEL=mean(.SD$X3D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
775 P3=mean(.SD$P3D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
776 N2=mean(rowSums(.SD[,c("N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
777 P4=mean(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
778 DEL.JH=mean(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
779 Total.Del=mean(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
780 Total.N=mean(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
781 Total.P=mean(rowSums(.SD[,c("P3V.nt.nb", "P5D.nt.nb", "P3D.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
782 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
783 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
784 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
785 write.table(newData, "junctionAnalysisProd_mean_wD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
786
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
787 newData = data.frame(data.table(PRODF.with.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
788 VH.DEL=num_median(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
789 P1=num_median(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
790 N1=num_median(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
791 P2=num_median(.SD$P5D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
792 DEL.DH=num_median(.SD$X5D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
793 DH.DEL=num_median(.SD$X3D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
794 P3=num_median(.SD$P3D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
795 N2=num_median(rowSums(.SD[,c("N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
796 P4=num_median(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
797 DEL.JH=num_median(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
798 Total.Del=num_median(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
799 Total.N=num_median(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
800 Total.P=num_median(rowSums(.SD[,c("P3V.nt.nb", "P5D.nt.nb", "P3D.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
801 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
802 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
803 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
804 write.table(newData, "junctionAnalysisProd_median_wD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
805
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
806 newData = data.frame(data.table(UNPROD.with.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
807 VH.DEL=mean(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
808 P1=mean(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
809 N1=mean(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
810 P2=mean(.SD$P5D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
811 DEL.DH=mean(.SD$X5D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
812 DH.DEL=mean(.SD$X3D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
813 P3=mean(.SD$P3D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
814 N2=mean(rowSums(.SD[,c("N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
815 P4=mean(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
816 DEL.JH=mean(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
817 Total.Del=mean(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
818 Total.N=mean(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
819 Total.P=mean(rowSums(.SD[,c("P3V.nt.nb", "P5D.nt.nb", "P3D.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
820 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
821 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
822 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
823 write.table(newData, "junctionAnalysisUnProd_mean_wD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
824
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
825 newData = data.frame(data.table(UNPROD.with.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
826 VH.DEL=num_median(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
827 P1=num_median(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
828 N1=num_median(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
829 P2=num_median(.SD$P5D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
830 DEL.DH=num_median(.SD$X5D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
831 DH.DEL=num_median(.SD$X3D.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
832 P3=num_median(.SD$P3D.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
833 N2=num_median(rowSums(.SD[,c("N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
834 P4=num_median(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
835 DEL.JH=num_median(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
836 Total.Del=num_median(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5D.REGION.trimmed.nt.nb", "X3D.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
837 Total.N=num_median(rowSums(.SD[,c("N.REGION.nt.nb", "N1.REGION.nt.nb", "N2.REGION.nt.nb", "N3.REGION.nt.nb", "N4.REGION.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
838 Total.P=num_median(rowSums(.SD[,c("P3V.nt.nb", "P5D.nt.nb", "P3D.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
839 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
840 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
841 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
842 write.table(newData, "junctionAnalysisUnProd_median_wD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
843
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
844 #---------------- again for no-D
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
845
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
846 newData = data.frame(data.table(PRODF.no.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
847 VH.DEL=mean(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
848 P1=mean(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
849 N1=mean(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
850 P2=mean(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
851 DEL.JH=mean(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
852 Total.Del=mean(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
853 Total.N=mean(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
854 Total.P=mean(rowSums(.SD[,c("P3V.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
855 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
856 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
857 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
858 write.table(newData, "junctionAnalysisProd_mean_nD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
859
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
860 newData = data.frame(data.table(PRODF.no.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
861 VH.DEL=num_median(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
862 P1=num_median(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
863 N1=num_median(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
864 P2=num_median(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
865 DEL.JH=num_median(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
866 Total.Del=num_median(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
867 Total.N=num_median(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
868 Total.P=num_median(rowSums(.SD[,c("P3V.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
869 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
870 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
871 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
872 write.table(newData, "junctionAnalysisProd_median_nD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
873
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
874 newData = data.frame(data.table(UNPROD.no.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
875 VH.DEL=mean(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
876 P1=mean(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
877 N1=mean(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
878 P2=mean(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
879 DEL.JH=mean(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
880 Total.Del=mean(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
881 Total.N=mean(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
882 Total.P=mean(rowSums(.SD[,c("P3V.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
883 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
884 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
885 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
886 write.table(newData, "junctionAnalysisUnProd_mean_nD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
887
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
888
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
889 newData = data.frame(data.table(UNPROD.no.D)[,list(unique=.N,
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
890 VH.DEL=num_median(.SD$X3V.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
891 P1=num_median(.SD$P3V.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
892 N1=num_median(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
893 P2=num_median(.SD$P5J.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
894 DEL.JH=num_median(.SD$X5J.REGION.trimmed.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
895 Total.Del=num_median(rowSums(.SD[,c("X3V.REGION.trimmed.nt.nb", "X5J.REGION.trimmed.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
896 Total.N=num_median(.SD$N.REGION.nt.nb, na.rm=T),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
897 Total.P=num_median(rowSums(.SD[,c("P3V.nt.nb", "P5J.nt.nb"), with=F], na.rm=T)),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
898 Median.CDR3.l=as.double(median(as.numeric(.SD$CDR3.Length), na.rm=T))),
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
899 by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
900 newData[,sapply(newData, is.numeric)] = round(newData[,sapply(newData, is.numeric)],1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
901 write.table(newData, "junctionAnalysisUnProd_median_nD.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
902 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
903
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
904 PRODF = bak
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
905 UNPROD = bakun
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
906
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
907
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
908 # ---------------------- D reading frame ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
909
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
910 D.REGION.reading.frame = PRODF[,c("Sample", "D.REGION.reading.frame")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
911
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
912 chck = is.na(D.REGION.reading.frame$D.REGION.reading.frame)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
913 if(any(chck)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
914 D.REGION.reading.frame[chck,"D.REGION.reading.frame"] = "No D"
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
915 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
916
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
917 D.REGION.reading.frame.1 = data.frame(data.table(D.REGION.reading.frame)[, list(Freq=.N), by=c("Sample", "D.REGION.reading.frame")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
918
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
919 D.REGION.reading.frame.2 = data.frame(data.table(D.REGION.reading.frame)[, list(sample.sum=sum(as.numeric(.SD$D.REGION.reading.frame), na.rm=T)), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
920
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
921 D.REGION.reading.frame = merge(D.REGION.reading.frame.1, D.REGION.reading.frame.2, by="Sample")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
922
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
923 D.REGION.reading.frame$percentage = round(D.REGION.reading.frame$Freq / D.REGION.reading.frame$sample.sum * 100, 1)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
924
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
925 write.table(D.REGION.reading.frame, "DReadingFrame.txt" , sep="\t",quote=F,row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
926
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
927 D.REGION.reading.frame = ggplot(D.REGION.reading.frame)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
928 D.REGION.reading.frame = D.REGION.reading.frame + geom_bar(aes( x = D.REGION.reading.frame, y = percentage, fill=Sample), stat='identity', position='dodge' ) + ggtitle("D reading frame") + xlab("Frame") + ylab("Frequency")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
929 D.REGION.reading.frame = D.REGION.reading.frame + scale_fill_manual(values=sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
930 D.REGION.reading.frame = D.REGION.reading.frame + theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), axis.text.x = element_text(angle = 45, hjust = 1), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
931
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
932 png("DReadingFrame.png")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
933 D.REGION.reading.frame
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
934 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
935
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
936 ggsave("DReadingFrame.pdf", D.REGION.reading.frame)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
937
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
938 # ---------------------- AA composition in CDR3 ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
939
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
940 AACDR3 = PRODF[,c("Sample", "CDR3.Seq")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
941
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
942 TotalPerSample = data.frame(data.table(AACDR3)[, list(total=sum(nchar(as.character(.SD$CDR3.Seq)))), by=Sample])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
943
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
944 AAfreq = list()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
945
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
946 for(i in 1:nrow(TotalPerSample)){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
947 sample = TotalPerSample$Sample[i]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
948 AAfreq[[i]] = data.frame(table(unlist(strsplit(as.character(AACDR3[AACDR3$Sample == sample,c("CDR3.Seq")]), ""))))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
949 AAfreq[[i]]$Sample = sample
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
950 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
951
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
952 AAfreq = ldply(AAfreq, data.frame)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
953 AAfreq = merge(AAfreq, TotalPerSample, by="Sample", all.x = T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
954 AAfreq$freq_perc = as.numeric(AAfreq$Freq / AAfreq$total * 100)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
955
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
956
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
957 AAorder = read.table(sep="\t", header=TRUE, text="order.aa\tAA\n1\tR\n2\tK\n3\tN\n4\tD\n5\tQ\n6\tE\n7\tH\n8\tP\n9\tY\n10\tW\n11\tS\n12\tT\n13\tG\n14\tA\n15\tM\n16\tC\n17\tF\n18\tL\n19\tV\n20\tI")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
958 AAfreq = merge(AAfreq, AAorder, by.x='Var1', by.y='AA', all.x=TRUE)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
959
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
960 AAfreq = AAfreq[!is.na(AAfreq$order.aa),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
961
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
962 AAfreqplot = ggplot(AAfreq)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
963 AAfreqplot = AAfreqplot + geom_bar(aes( x=factor(reorder(Var1, order.aa)), y = freq_perc, fill = Sample), stat='identity', position='dodge' )
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
964 AAfreqplot = AAfreqplot + annotate("rect", xmin = 0.5, xmax = 2.5, ymin = 0, ymax = Inf, fill = "red", alpha = 0.2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
965 AAfreqplot = AAfreqplot + annotate("rect", xmin = 3.5, xmax = 4.5, ymin = 0, ymax = Inf, fill = "blue", alpha = 0.2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
966 AAfreqplot = AAfreqplot + annotate("rect", xmin = 5.5, xmax = 6.5, ymin = 0, ymax = Inf, fill = "blue", alpha = 0.2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
967 AAfreqplot = AAfreqplot + annotate("rect", xmin = 6.5, xmax = 7.5, ymin = 0, ymax = Inf, fill = "red", alpha = 0.2)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
968 AAfreqplot = AAfreqplot + ggtitle("Amino Acid Composition in the CDR3") + xlab("Amino Acid, from Hydrophilic (left) to Hydrophobic (right)") + ylab("Percentage") + scale_fill_manual(values=sample.colors)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
969 AAfreqplot = AAfreqplot + theme(panel.background = element_rect(fill = "white", colour="black"),text = element_text(size=15, colour="black"), panel.grid.major.y = element_line(colour = "black"), panel.grid.major.x = element_blank())
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
970
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
971 png("AAComposition.png",width = 1280, height = 720)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
972 AAfreqplot
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
973 dev.off()
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
974
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
975 ggsave("AAComposition.pdf", AAfreqplot, width=12, height=7)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
976
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
977 write.table(AAfreq, "AAComposition.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
978
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
979 # ---------------------- AA median CDR3 length ----------------------
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
980
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
981 median.aa.l = data.frame(data.table(PRODF)[, list(median=as.double(median(as.numeric(.SD$CDR3.Length, na.rm=T), na.rm=T))), by=c("Sample")])
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
982 write.table(median.aa.l, "AAMedianBySample.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=F)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
983
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
984 if(clonaltype != "none"){
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
985 #generate the "Sequences that are present in more than one replicate" dataset
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
986 clonaltype.in.replicates = inputdata
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
987 clonaltype.in.replicates = clonaltype.in.replicates[clonaltype.in.replicates$Functionality %in% c("productive (see comment)","productive"),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
988 clonaltype.in.replicates = clonaltype.in.replicates[!(is.na(clonaltype.in.replicates$ID) | is.na(clonaltype.in.replicates$Top.V.Gene) | is.na(clonaltype.in.replicates$Top.J.Gene)),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
989 clonaltype = unlist(strsplit(clonaltype, ","))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
990
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
991 clonaltype.in.replicates$clonaltype = do.call(paste, c(clonaltype.in.replicates[clonaltype], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
992
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
993 clonaltype.in.replicates = clonaltype.in.replicates[!duplicated(clonaltype.in.replicates$clonaltype),]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
994
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
995 clonaltype = clonaltype[-which(clonaltype == "Sample")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
996
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
997 clonaltype.in.replicates$clonaltype = do.call(paste, c(clonaltype.in.replicates[clonaltype], sep = ":"))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
998 clonaltype.in.replicates = clonaltype.in.replicates[,c("clonaltype","Replicate", "ID", "Sequence", "Sample")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
999
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1000
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1001 write.table(clonaltype.in.replicates, "clonaltypes_replicates_before_table.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1002
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1003 clonaltype.counts = data.frame(table(clonaltype.in.replicates$clonaltype))
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1004
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1005 write.table(clonaltype.counts, "clonaltypes_counts.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1006
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1007 names(clonaltype.counts) = c("clonaltype", "coincidence")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1008
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1009 clonaltype.counts = clonaltype.counts[clonaltype.counts$coincidence > 1,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1010
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1011 clonaltype.in.replicates = clonaltype.in.replicates[clonaltype.in.replicates$clonaltype %in% clonaltype.counts$clonaltype,]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1012 clonaltype.in.replicates = merge(clonaltype.in.replicates, clonaltype.counts, by="clonaltype")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1013 clonaltype.in.replicates = clonaltype.in.replicates[order(-clonaltype.in.replicates$coincidence, clonaltype.in.replicates$clonaltype, clonaltype.in.replicates$Replicate),c("coincidence","clonaltype", "Sample", "Replicate", "ID", "Sequence")]
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1014
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1015
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1016 write.table(clonaltype.in.replicates, "clonaltypes_replicates.txt" , sep="\t",quote=F,na="-",row.names=F,col.names=T)
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1017 } else {
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1018 cat("No clonaltype", file="clonaltypes_replicates_before_table.txt")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1019 cat("No clonaltype", file="clonaltypes_counts.txt")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1020 cat("No clonaltype", file="clonaltypes_replicates.txt")
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1021 }
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1022
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1023
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1024
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1025
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1026
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1027
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1028
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1029
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1030
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1031
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1032
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1033
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1034
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1035
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1036
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1037
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1038
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1039
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1040
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1041
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1042
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1043
81b3eb11ed2c "planemo upload commit 6c0195cc4de6a34e2c46d875be4fc2157a21cdf6"
rhpvorderman
parents: 52
diff changeset
1044