comparison sequence_overview.r @ 0:8a5a2abbb870 draft default tip

Uploaded

0:8a5a2abbb870

library(reshape2)

args <- commandArgs(trailingOnly = TRUE)

before.unique.file = args[1]
merged.file = args[2]
outputdir = args[3]
gene.classes = unlist(strsplit(args[4], ","))
hotspot.analysis.sum.file = args[5]
NToverview.file = paste(outputdir, "ntoverview.txt", sep="/")
NTsum.file = paste(outputdir, "ntsum.txt", sep="/")
main.html = "index.html"

setwd(outputdir)

before.unique = read.table(before.unique.file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
merged = read.table(merged.file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
hotspot.analysis.sum = read.table(hotspot.analysis.sum.file, header=F, sep=",", fill=T, stringsAsFactors=F, quote="")

#before.unique = before.unique[!grepl("unmatched", before.unique$best_match),]

before.unique$seq_conc = paste(before.unique$CDR1.IMGT.seq, before.unique$FR2.IMGT.seq, before.unique$CDR2.IMGT.seq, before.unique$FR3.IMGT.seq, before.unique$CDR3.IMGT.seq)

IDs = before.unique[,c("Sequence.ID", "seq_conc", "best_match", "Functionality")]
IDs$best_match = as.character(IDs$best_match)

#dat = data.frame(data.table(dat)[, list(freq=.N), by=c("best_match", "seq_conc")])

dat = data.frame(table(before.unique$seq_conc))
#dat = data.frame(table(merged$seq_conc, merged$Functionality))

#dat = dat[dat$Freq > 1,]

#names(dat) = c("seq_conc", "Functionality", "Freq")
names(dat) = c("seq_conc", "Freq")

dat$seq_conc = factor(dat$seq_conc)

dat = dat[order(as.character(dat$seq_conc)),]

#writing html from R...
get.bg.color = function(val){
	if(val %in% c("TRUE", "FALSE", "T", "F")){ #if its a logical value, give the background a green/red color
		return(ifelse(val,"#eafaf1","#f9ebea"))
	} else if (!is.na(as.numeric(val))) { #if its a numerical value, give it a grey tint if its >0
		return(ifelse(val > 0,"#eaecee","white"))
	} else {
		return("white")
	}
}
td = function(val) {
  return(paste("<td bgcolor='", get.bg.color(val), "'>", val, "</td>", sep=""))
}
tr = function(val) { 
	return(paste(c("<tr>", sapply(val, td), "</tr>"), collapse="")) 
}

make.link = function(id, clss, val) { 
	paste("<a href='", clss, "_", id, ".html'>", val, "</a>", sep="") 
}
tbl = function(df) {
	res = "<table border='1'>"
	for(i in 1:nrow(df)){ 
		res = paste(res, tr(df[i,]), sep="")
	}
	res = paste(res, "</table>")
}

cat("<table border='1'>", file=main.html, append=F)
cat("<caption>CDR1+FR2+CDR2+FR3+CDR3 sequences that show up more than once</caption>", file=main.html, append=T)
cat("<tr>", file=main.html, append=T)
cat("<th>Sequence</th><th>Functionality</th><th>ca1</th><th>ca2</th><th>cg1</th><th>cg2</th><th>cg3</th><th>cg4</th><th>cm</th><th>un</th>", file=main.html, append=T)
cat("<th>total CA</th><th>total CG</th><th>number of subclasses</th><th>present in both Ca and Cg</th><th>Ca1+Ca2</th>", file=main.html, append=T)
cat("<th>Cg1+Cg2</th><th>Cg1+Cg3</th><th>Cg1+Cg4</th><th>Cg2+Cg3</th><th>Cg2+Cg4</th><th>Cg3+Cg4</th>", file=main.html, append=T)
cat("<th>Cg1+Cg2+Cg3</th><th>Cg2+Cg3+Cg4</th><th>Cg1+Cg2+Cg4</th><th>Cg1+Cg3+Cg4</th><th>Cg1+Cg2+Cg3+Cg4</th>", file=main.html, append=T)
cat("</tr>", file=main.html, append=T)



single.sequences=0 #sequence only found once, skipped
in.multiple=0 #same sequence across multiple subclasses
multiple.in.one=0 #same sequence multiple times in one subclass
unmatched=0 #all of the sequences are unmatched
some.unmatched=0 #one or more sequences in a clone are unmatched
matched=0 #should be the same als matched sequences

sequence.id.page="by_id.html"

for(i in 1:nrow(dat)){
	
	ca1 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^ca1", IDs$best_match),]
	ca2 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^ca2", IDs$best_match),]
	
	cg1 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^cg1", IDs$best_match),]
	cg2 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^cg2", IDs$best_match),]
	cg3 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^cg3", IDs$best_match),]
	cg4 = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^cg4", IDs$best_match),]
	
	cm = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^cm", IDs$best_match),]
	
	un = IDs[IDs$seq_conc == dat[i,c("seq_conc")] & grepl("^unmatched", IDs$best_match),]
	allc = rbind(ca1, ca2, cg1, cg2, cg3, cg4, cm, un)
	
	ca1.n = nrow(ca1)
	ca2.n = nrow(ca2)
	
	cg1.n = nrow(cg1)
	cg2.n = nrow(cg2)
	cg3.n = nrow(cg3)
	cg4.n = nrow(cg4)
	
	cm.n = nrow(cm)
	
	un.n = nrow(un)
	
	classes = c(ca1.n, ca2.n, cg1.n, cg2.n, cg3.n, cg4.n, cm.n, un.n)
	
	classes.sum = sum(classes)
	
	if(classes.sum == 1){
		single.sequences = single.sequences + 1
		next
	}
	
	if(un.n == classes.sum){
		unmatched = unmatched + 1
		next
	}
	
	in.classes = sum(classes > 0)
	
	matched = matched + in.classes #count in how many subclasses the sequence occurs.
	
	if(any(classes  == classes.sum)){
		multiple.in.one = multiple.in.one + 1
	} else if (un.n > 0) {
		some.unmatched = some.unmatched + 1
	} else {
		in.multiple = in.multiple + 1
	}
	
	id = as.numeric(dat[i,"seq_conc"])
	
	functionality = paste(unique(allc[,"Functionality"]), collapse=",")
	
	by.id.row = c()
	
	if(ca1.n > 0){
		cat(tbl(ca1), file=paste("ca1_", id, ".html", sep=""))
	}

	if(ca2.n > 0){
		cat(tbl(ca2), file=paste("ca2_", id, ".html", sep=""))
	}

	if(cg1.n > 0){
		cat(tbl(cg1), file=paste("cg1_", id, ".html", sep=""))
	}

	if(cg2.n > 0){
		cat(tbl(cg2), file=paste("cg2_", id, ".html", sep=""))
	}

	if(cg3.n > 0){
		cat(tbl(cg3), file=paste("cg3_", id, ".html", sep=""))
	}

	if(cg4.n > 0){
		cat(tbl(cg4), file=paste("cg4_", id, ".html", sep=""))
	}

	if(cm.n > 0){
		cat(tbl(cm), file=paste("cm_", id, ".html", sep=""))
	}

	if(un.n > 0){
		cat(tbl(un), file=paste("un_", id, ".html", sep=""))
	}
	
	ca1.html = make.link(id, "ca1", ca1.n)
	ca2.html = make.link(id, "ca2", ca2.n)
	
	cg1.html = make.link(id, "cg1", cg1.n)
	cg2.html = make.link(id, "cg2", cg2.n)
	cg3.html = make.link(id, "cg3", cg3.n)
	cg4.html = make.link(id, "cg4", cg4.n)
	
	cm.html = make.link(id, "cm", cm.n)
	
	un.html = make.link(id, "un", un.n)
	
	#extra columns
	ca.n = ca1.n + ca2.n
	
	cg.n = cg1.n + cg2.n + cg3.n + cg4.n
	
	#in.classes
	
	in.ca.cg = (ca.n > 0 & cg.n > 0)
	
	in.ca1.ca2 = (ca1.n > 0 & ca2.n > 0)
	
	in.cg1.cg2 = (cg1.n > 0 & cg2.n > 0)
	in.cg1.cg3 = (cg1.n > 0 & cg3.n > 0)
	in.cg1.cg4 = (cg1.n > 0 & cg4.n > 0)
	in.cg2.cg3 = (cg2.n > 0 & cg3.n > 0)
	in.cg2.cg4 = (cg2.n > 0 & cg4.n > 0)
	in.cg3.cg4 = (cg3.n > 0 & cg4.n > 0)
	
	in.cg1.cg2.cg3 = (cg1.n > 0 & cg2.n > 0 & cg3.n > 0)
	in.cg2.cg3.cg4 = (cg2.n > 0 & cg3.n > 0 & cg4.n > 0)
	in.cg1.cg2.cg4 = (cg1.n > 0 & cg2.n > 0 & cg4.n > 0)
	in.cg1.cg3.cg4 = (cg1.n > 0 & cg3.n > 0 & cg4.n > 0)
	
	in.cg.all = (cg1.n > 0 & cg2.n > 0 & cg3.n > 0 & cg4.n > 0)
	
	
	
	
	#rw = c(as.character(dat[i,"seq_conc"]), functionality, ca1.html, ca2.html, cg1.html, cg2.html, cg3.html, cg4.html, cm.html, un.html)
	rw = c(as.character(dat[i,"seq_conc"]), functionality, ca1.html, ca2.html, cg1.html, cg2.html, cg3.html, cg4.html, cm.html, un.html)
	rw = c(rw, ca.n, cg.n, in.classes, in.ca.cg, in.ca1.ca2, in.cg1.cg2, in.cg1.cg3, in.cg1.cg4, in.cg2.cg3, in.cg2.cg4, in.cg3.cg4, in.cg1.cg2.cg3, in.cg2.cg3.cg4, in.cg1.cg2.cg4, in.cg1.cg3.cg4, in.cg.all)

	cat(tr(rw), file=main.html, append=T)
	
	
	for(i in 1:nrow(allc)){ #generate html by id
		html = make.link(id, allc[i,"best_match"], allc[i,"Sequence.ID"])
		cat(paste(html, "<br />"), file=sequence.id.page, append=T)
	}
}

cat("</table>", file=main.html, append=T)

print(paste("Single sequences:", single.sequences))
print(paste("Sequences in multiple subclasses:", in.multiple))
print(paste("Multiple sequences in one subclass:", multiple.in.one))
print(paste("Matched with unmatched:", some.unmatched))
print(paste("Count that should match 'matched' sequences:", matched))

#ACGT overview

NToverview = merged[!grepl("^unmatched", merged$best_match),]

NToverview$seq = paste(NToverview$CDR1.IMGT.seq, NToverview$FR2.IMGT.seq, NToverview$CDR2.IMGT.seq, NToverview$FR3.IMGT.seq, sep="_")

NToverview$A = nchar(gsub("[^Aa]", "", NToverview$seq))
NToverview$C = nchar(gsub("[^Cc]", "", NToverview$seq))
NToverview$G = nchar(gsub("[^Gg]", "", NToverview$seq))
NToverview$T = nchar(gsub("[^Tt]", "", NToverview$seq))

#Nsum = data.frame(Sequence.ID="-", best_match="Sum", seq="-", A = sum(NToverview$A), C = sum(NToverview$C), G = sum(NToverview$G), T = sum(NToverview$T))

#NToverview = rbind(NToverview, NTsum)

NTresult = data.frame(nt=c("A", "C", "T", "G"))

for(clazz in gene.classes){
	NToverview.sub = NToverview[grepl(paste("^", clazz, sep=""), NToverview$best_match),]
	new.col.x = c(sum(NToverview.sub$A), sum(NToverview.sub$C), sum(NToverview.sub$T), sum(NToverview.sub$G))
	new.col.y = sum(new.col.x)
	new.col.z = round(new.col.x / new.col.y * 100, 2)
	
	tmp = names(NTresult)
	NTresult = cbind(NTresult, data.frame(new.col.x, new.col.y, new.col.z))
	names(NTresult) = c(tmp, paste(clazz, c("x", "y", "z"), sep=""))
}

write.table(NToverview[,c("Sequence.ID", "best_match", "seq", "A", "C", "G", "T")], NToverview.file, quote=F, sep="\t", row.names=F, col.names=T)

NToverview = NToverview[!grepl("unmatched", NToverview$best_match),]

new.col.x = c(sum(NToverview$A), sum(NToverview$C), sum(NToverview$T), sum(NToverview$G))
new.col.y = sum(new.col.x)
new.col.z = round(new.col.x / new.col.y * 100, 2)

tmp = names(NTresult)
NTresult = cbind(NTresult, data.frame(new.col.x, new.col.y, new.col.z))
names(NTresult) = c(tmp, paste("all", c("x", "y", "z"), sep=""))

names(hotspot.analysis.sum) = names(NTresult)

hotspot.analysis.sum = rbind(hotspot.analysis.sum, NTresult)

write.table(hotspot.analysis.sum, hotspot.analysis.sum.file, quote=F, sep=",", row.names=F, col.names=F, na="0")