view Dotplot_Release/R_dotPlot_hc.R @ 5:dc2aed283637 draft

Uploaded
author bornea
date Fri, 29 Jan 2016 10:14:36 -0500
parents dfa3436beb67
children
line wrap: on
line source

#!/usr/bin/env Rscript

args <- commandArgs(trailingOnly = TRUE)

pheatmapj_loc <- paste(args[6],"pheatmap_j.R",sep="/")
heatmap2j_loc <- paste(args[6],"heatmap_2j.R",sep="/")

library('latticeExtra')
library('RColorBrewer')
library('grid')
library(reshape2)
library('gplots')
library('gtools')
source(pheatmapj_loc)
source(heatmap2j_loc)

data.file <- read.table("SC_data.txt", sep="\t", header=TRUE, row.names=1) ### import spectral count data
data.file2 <- read.table("FDR_data.txt", sep="\t", header=TRUE, row.names=1) ### import FDR count data

#setting parameters

Sfirst=as.numeric(args[1]) #first FDR cutoff
Ssecond=as.numeric(args[2]) #second FDR cutoff
maxp=as.integer(args[3]) #maximum value for a spectral count
methd <- args[4]
dist_methd <- args[5]

#determine bait and prey ordering

dist_bait <- dist(as.matrix(t(data.file)), method= dist_methd) # "euclidean", "maximum", "manhattan", "canberra", "binary" or "minkowski"
dist_prey <- dist(as.matrix(data.file), method= dist_methd)

if(methd == "ward"){
	dist_bait <- dist_bait^2 #comment out this line and the next if not using Ward's method of clustering
	dist_prey <- dist_prey^2
}

hc_bait <- hclust(dist_bait, method = methd) # method = "average", "single", "complete", "ward", "mcquitty", "median" or "centroid"
hc_prey <- hclust(dist_prey, method = methd)

data.file = data.file[hc_prey$order, , drop = FALSE]
data.file = data.file[, hc_bait$order, drop = FALSE]
data.file2 = data.file2[hc_prey$order, , drop = FALSE]
data.file2 = data.file2[, hc_bait$order, drop = FALSE]

x_ord=factor(row.names(data.file), levels=row.names(data.file))
y_ord=factor(names(data.file[1,]), levels=names(data.file[1,]))

df<-data.frame(y=rep(y_ord, nrow(data.file))
	,x=rep(x_ord, each=ncol(data.file))
	,z1=as.vector(t(data.file)) # Circle color
	,z2=as.vector(t(data.file/apply(data.file,1,max))) # Circle size
	,z3=as.vector(t(data.file2)) # FDR
)
	
df$z1[df$z1>maxp] <- maxp #maximum value for spectral count
df$z2[df$z2==0] <- NA
df$z3[df$z3>Ssecond] <- 0.05*maxp
df$z3[df$z3<=Ssecond & df$z3>Sfirst] <- 0.5*maxp
df$z3[df$z3<=Sfirst] <- 1*maxp
df$z4 <- df$z1
df$z4[df$z4==0] <- 0
df$z4[df$z4>0] <- 2.5 

# The labeling for the colorkey

labelat = c(0, maxp)
labeltext = c(0, maxp)

# color scheme to use

nmb.colors<-maxp
z.colors<-grey(rev(seq(0,0.9,0.9/nmb.colors))) #grayscale color scale

#plot dotplot

pl <- levelplot(z1~x*y, data=df
	,col.regions =z.colors #terrain.colors(100)
	,scales = list(x = list(rot = 90), y=list(cex=0.8), tck=0) # rotates X,Y labels and changes scale 
	,colorkey = FALSE
	#,colorkey = list(space="bottom", width=1.5, height=0.3, labels=list(at = labelat, labels = labeltext)) #put colorkey at top with my labeling scheme
	,xlab="Prey", ylab="Bait"
	,panel=function(x,y,z,...,col.regions){
		print(x)
		z.c<-df$z1[ (df$x %in% as.character(x)) & (df$y %in% y)]
		z.2<-df$z2[ (df$x %in% as.character(x)) & (df$y %in% y)]
		z.3<-df$z3
		z.4<-df$z4
		panel.xyplot(x,y
			,as.table=TRUE
			,pch=21 # point type to use (circles in this case)
			,cex=((z.2-min(z.2,na.rm=TRUE))/(max(z.2,na.rm=TRUE)-min(z.2,na.rm=TRUE)))*3 #circle size
			,fill=z.colors[floor((z.c-min(z.c,na.rm=TRUE))*nmb.colors/(max(z.c,na.rm=TRUE)-min(z.c,na.rm=TRUE)))+1] # circle colors
			,col=z.colors[1+z.3] # border colors
			,lex=z.4 #border thickness
			)
	}
	#,main="Fold change" # graph main title
	)
if(ncol(data.file) > 4) ht=3.5+(0.36*((ncol(data.file)-1)-4)) else ht=3.5
if(nrow(data.file) > 20) wd=8.25+(0.29*(nrow(data.file) -20)) else wd=5.7+(0.28*(nrow(data.file) -10))
pdf("dotplot.pdf", onefile = FALSE, paper = "special", height = ht, width = wd, pointsize = 2)
print(pl)
dev.off()

#plot bait vs prey heatmap

heat_df <- acast(df, y~x, value.var="z1")
heat_df <- apply(heat_df, 2, rev)

if(ncol(data.file) > 4) ht=3.5+(0.1*((ncol(data.file)-1)-4)) else ht=3.5
if(nrow(data.file) > 20) wd=8.25+(0.1*(nrow(data.file)-20)) else wd=5+(0.1*(nrow(data.file)-10))
pdf("heatmap_borders.pdf", onefile = FALSE, paper = "special", height = ht, width = wd, pointsize = 2)
pheatmap_j(heat_df, scale="none", border_color="black", border_width = 0.1, cluster_rows=FALSE, cluster_cols=FALSE, col=colorRampPalette(c("#FFFFFF", brewer.pal(9,"Blues")))(100))
dev.off()

pdf("heatmap_no_borders.pdf", onefile = FALSE, paper = "special", height = ht, width = wd, pointsize = 2)
pheatmap_j(heat_df, scale="none", border_color=NA, cluster_rows=FALSE, cluster_cols=FALSE, col=colorRampPalette(c("#FFFFFF", brewer.pal(9,"Blues")))(100))
dev.off()

#plot bait vs bait heatmap using dist matrix
dist_bait <- dist_bait/max(dist_bait)
pdf("bait2bait.pdf", onefile = FALSE, paper = "special")
heatmap_2j(as.matrix(dist_bait), trace="none", scale="none", density.info="none", col=rev(colorRampPalette(c("#FFFFFF", brewer.pal(9,"Blues")))(100)), xMin=0, xMax=1, margins=c(1.5*max(nchar(rownames(as.matrix(dist_bait)))),1.5*max(nchar(colnames(as.matrix(dist_bait))))))
dev.off()