scripts/old/plot_ranks_layout.R

In ahopki14/tcrSeqR: Tools for analyzing data from Adaptive's ImmunoSeq platform

library(scales)
#library(immunoSeqR)
load('~/Documents/emj/ImmunoseqResults/sampleExport.2014-07-31_10-10-24/ds_agg_reorder.Rda')
load('~/Documents/emj/ImmunoseqResults/sampleExport.2014-07-31_10-10-24/dict_reorder.Rda')

n <- 200
ds <- ds_agg_reorder
dict <- dict_reorder

pdf(paste0('~/Documents/emj/ImmunoseqResults/R/plots/Rankplots/all-',n,'.pdf'), width=8.5,height=8)
par(mfcol=c(3,2))
for(a in seq(length(levels(dict$patient)))){
	name <- levels(dict$patient)[a]
	resp <- as.character(dict$response[which(dict$patient==name)][1])
	samples <- c(2, # include the aa data
		which(dict$patient==name & dict$type=='Pre'),
		which(dict$patient==name & dict$type=='Post'),
		which(dict$patient==name & dict$type=='PDAC')
	)
	names(ds)[samples] 
	tds <- ds[,samples]
	tds <- tds[tds[,2]>0 | tds[,3]>0,]
	#tds <- tds[tds[,4]>0,] # restrict to tumor
	rownames(tds) <- seq_len(nrow(tds))
	tds$aa <- as.factor(as.character(tds$aa))
	o <- overlap(tds[,2],tds[,3]) 
	Pre <- rank(-tds[,2],ties.method='min') 
	Post <- rank(-tds[,3],ties.method='min')
	#plot(Pre,Post,log='xy',col=alpha('blue',0.2),pch=19,cex=0.75)
	pre_b <- which(Pre %in% 1:n)
	post_b <- which(Post %in% 1:n)
	desc <- c(rep("Pre",length(pre_b)),rep("Post",length(post_b)))
	desc <- as.factor(desc)
	desc <- factor(desc, levels=rev(levels(desc)))
	dat <- c(Pre[pre_b],Post[pre_b])

	stripchart( #makes an empty chart to draw lines on
		c(0,0)~factor(c('Pre','Post'),levels=c('Pre','Post')),
		at=c(1.25,1.75),
		ylim=c((3*n),1),
		xlim=c(1.2,1.8),
		vertical=TRUE,
		col='white',
		ylab='Rank',
		frame.plot=FALSE,
		main=paste0(name,'(',resp,')')#, '\n Present in Tumor')
		)
	for(a in 1:n){
		segments(1.2,Pre[pre_b][a],1.7,Post[pre_b][a],col=alpha('red',0.4),lend=0,lwd=0.1)
		segments(1.3,Pre[post_b][a],1.8,Post[post_b][a],col=alpha('blue',0.4),lend=0,lwd=0.1)
	}
}
dev.off()