comparison [APliBio]Nebula tools suite/Nebula/AnnotateGenes/geneDist.R @ 0:2ec3ba0e9e70 draft

Uploaded
author alermine
date Thu, 25 Oct 2012 08:18:25 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:2ec3ba0e9e70
1 args <- commandArgs()
2 input <- args[4]
3 pngFile <- args[5]
4 dataTable <-read.table(file=input, header=TRUE);
5 chip.data<-data.frame(dataTable)
6 ifReg <- 0
7 if (length(unique(chip.data$Reg))>1) {
8 ifReg <- 1
9 }
10
11 ifPDF <- 0
12 bootstrap <- 1
13
14 if (length(args)>=8) {
15 ifPDF=args[8]
16 bootstrap=args[9]
17 }
18 if (length(args)==7 & args[7]==1) {
19 ifPDF=1
20 }
21
22 ifControl <- 0
23 if (length(args)>=7 & args[7]!=1 & args[7]!=0) {
24 dataTable <-read.table(file=args[7], header=TRUE);
25 control.data<-data.frame(dataTable)
26 ifControl <- 1
27
28
29 }
30
31 error.bar <- function(x, y, upper, lower=upper, length=0.1,...){
32 if(length(x) != length(y) | length(y) !=length(lower) | length(lower) != length(upper))
33 stop("vectors must be same length")
34 arrows(x,y+upper, x, y-lower, angle=90, code=3, length=length, ...)
35 }
36
37
38
39 logFile <- args[6]
40 sink(logFile, append=FALSE, split=FALSE)
41
42 if (ifReg & ifControl) {
43
44 if (ifPDF==1) {
45 pdf(file = pngFile, width = 14, height = 13, pointsize = 20, bg = "white")
46 } else {
47 png(filename = pngFile, width = 1140, height = 840, units = "px", pointsize = 20, bg = "white", res = NA)
48 plot(1:10)
49 }
50 op <- par(mfrow = c(3,1))
51 } else {
52 if (ifReg | ifControl) {
53
54 if (ifPDF==1) {
55 pdf(file = pngFile, width = 20, height = 17, pointsize = 20, bg = "white")
56 } else {
57 png(filename = pngFile, width = 1580, height = 1230, units = "px", pointsize = 20, bg = "white", res = NA)
58 plot(1:10)
59 }
60 op <- par(mfrow = c(2,1))
61 } else {
62 if (ifPDF==1) {
63 pdf(file = pngFile, width = 22, height = 8, pointsize = 20, bg = "white")
64 } else {
65 png(filename = pngFile, width = 1580, height = 530, units = "px", pointsize = 20, bg = "white", res = NA)
66 plot(1:10)
67 }
68 op <- par(mfrow = c(1,1))
69 }
70 }
71 myColor <- 1
72 myColor[1] <- colors()[131]
73 myColor[2] <- colors()[59]
74 myColor[3] <- colors()[76]
75 myColor[4] <- colors()[88]
76 myColor[5] <- colors()[17]
77 myColor[6] <- colors()[565]
78 myColor[7] <- colors()[454]
79 myColor[8] <- colors()[401]
80 myColor[9] <- colors()[99]
81 myColorControl <- 1
82 myColorControl[1] <- colors()[24]
83 myColorControl[2] <- colors()[278]
84 myColorControl[3] <- colors()[305]
85 myColorControl[4] <- colors()[219]
86 myColorControl[5] <- colors()[343]
87 myColorControl[6] <- colors()[245]
88 myLevels <- 0
89 nn <- colnames(chip.data)
90 cc <- c(1:41)
91 colnames(chip.data) <- cc
92 countTotal <- length(unique(chip.data$"1"))
93 tt <- which(chip.data$"16">0)
94 countTotalEnh <- length(unique(chip.data$"1"[tt]))
95 tt <- which(chip.data$"10">0)
96 countTotalProm <- length(unique(chip.data$"1"[tt]))
97 tt <- which(chip.data$"12">0)
98 countTotalImDown <- length(unique(chip.data$"1"[tt]))
99 tt <- which(chip.data$"18">0)
100 countTotalIntra <- length(unique(chip.data$"1"[tt]))
101 tt <- which(chip.data$"20">0)
102 countTotal5kbD <- length(unique(chip.data$"1"[tt]))
103 tt <- which(chip.data$"28">0)
104 countTotal1IntronI <- length(unique(chip.data$"1"[tt]))
105 tt <- which(chip.data$"32">0)
106 countTotalExonsI <- length(unique(chip.data$"1"[tt]))
107 tt <- which(chip.data$"36">0)
108 countTotalIntronsI <- length(unique(chip.data$"1"[tt]))
109 tt <- which(chip.data$"40">0)
110 countTotalJonctionsI <- length(unique(chip.data$"1"[tt]))
111
112 names <- c("Enh.","Prom.","Imm.Down.","Intrag.","GeneDown.","F.Intron","Exons","2,3,etc.Introns","E.I.Junctions")
113 yChIPTotal <- c (countTotalEnh/countTotal,countTotalProm/countTotal, countTotalImDown/countTotal,countTotalIntra/countTotal,countTotal5kbD/countTotal,countTotal1IntronI/countTotal,countTotalExonsI/countTotal,countTotalIntronsI/countTotal,countTotalJonctionsI/countTotal)
114 if(!ifReg&!ifControl) {
115 par(mar=c(5.1, 7.1, 4.1, 2.1))
116 barplot(yChIPTotal,xlab="",beside=TRUE, col=c(myColor), names.arg=c(names),ylab="Proportion of genes with a peak")
117 cat ("Proportion of genes with a peak in a given genomic region:\n")
118 cat (paste(c(names),sep='\t'))
119 cat("\n")
120 cat (paste(c(yChIPTotal),sep='\t'))
121 cat("\n\n")
122 }
123 if (ifControl) {
124 if (bootstrap>1) {
125 yControlTotalMatrix <- NULL
126 }
127 for (fileNumber in 1:bootstrap) {
128
129 if (fileNumber>=2) {
130 dataTable <-read.table(file=paste(args[7],fileNumber,sep=""), header=TRUE);
131 control.data<-data.frame(dataTable)
132 }
133
134 colnames(control.data) <- cc
135 countTotalCntr <- length(unique(control.data$"1"))
136 tt <- which(control.data$"16">0)
137 countTotalEnhCntr <- length(unique(control.data$"1"[tt]))
138 tt <- which(control.data$"10">0)
139 countTotalPromCntr <- length(unique(control.data$"1"[tt]))
140 tt <- which(control.data$"12">0)
141 countTotalImDownCntr <- length(unique(control.data$"1"[tt]))
142 tt <- which(control.data$"18">0)
143 countTotalIntraCntr <- length(unique(control.data$"1"[tt]))
144 tt <- which(control.data$"20">0)
145 countTotal5kbDCntr <- length(unique(control.data$"1"[tt]))
146 tt <- which(control.data$"28">0)
147 countTotal1IntronICntr <- length(unique(control.data$"1"[tt]))
148 tt <- which(control.data$"32">0)
149 countTotalExonsICntr <- length(unique(control.data$"1"[tt]))
150 tt <- which(control.data$"36">0)
151 countTotalIntronsICntr <- length(unique(control.data$"1"[tt]))
152 tt <- which(control.data$"40">0)
153 countTotalJonctionsICntr <- length(unique(control.data$"1"[tt]))
154 yControlTotal <- c (countTotalEnhCntr/countTotalCntr,countTotalPromCntr/countTotalCntr, countTotalImDownCntr/countTotalCntr,countTotalIntraCntr/countTotalCntr,countTotal5kbDCntr/countTotalCntr,countTotal1IntronICntr/countTotalCntr,countTotalExonsICntr/countTotalCntr,countTotalIntronsICntr/countTotalCntr,countTotalJonctionsICntr/countTotalCntr)
155 if (bootstrap>1) {
156 yControlTotalMatrix <- rbind(yControlTotalMatrix,yControlTotal)
157 }
158 }
159 if (bootstrap>1) {
160 yControlTotal <- colMeans(yControlTotalMatrix)
161 Nrows <- nrow(yControlTotalMatrix)
162 colVars <- Nrows/(Nrows-1) * (colMeans(yControlTotalMatrix*yControlTotalMatrix)-colMeans(yControlTotalMatrix)^2)
163 }
164 if (!ifReg) {
165 cum = matrix( 0, nrow=2, ncol=length(names), byrow = TRUE)
166 for (i in c(1:length(names))) {
167 cum[1,i] <- yChIPTotal[i]
168 cum[2,i] <- yControlTotal[i]
169 }
170 if (bootstrap>1) {
171 wiskers <- matrix(c(colVars-colVars,sqrt(colVars)),2,length(names),byrow=TRUE)*1.96
172 }
173 par(mar=c(5.1, 7.1, 4.1, 2.1))
174 barx <- barplot(cum,xlab="",beside=TRUE, col=c(myColor[6],myColor[5]), legend = c("ChIP","Control"), names.arg=c(names),ylab="Proportion of genes with a peak")
175 cat ("Proportion of genes with a peak from the ChIP dataset in a given genomic region:\n")
176 cat (paste(c(names),sep='\t'))
177 cat("\n")
178 cat (paste(c(yChIPTotal),sep='\t'))
179 cat("\n\n")
180 cat ("Proportion of genes with a peak from the Control dataset in a given genomic region:\n")
181 cat (paste(c(names),sep='\t'))
182 cat("\n")
183 cat (paste(c(yControlTotal),sep='\t'))
184 cat("\n\n")
185 if (bootstrap>1) {
186 error.bar(barx,cum,wiskers)
187 cat ("Standard deviation for the Control dataset in a given genomic region:\n")
188 cat (paste(c(names),sep='\t'))
189 cat("\n")
190 cat (paste(c(sqrt(colVars)),sep='\t'))
191 cat("\n\n")
192 }
193 enrich <- NULL
194 for (i in c(1:length(names))) {
195 enrich[i] <- yChIPTotal[i]/yControlTotal[i];
196 }
197 barplot(enrich-1,xlab="",beside=TRUE, col=c(myColor), names.arg=c(names),ylab="Enrichment in comparison\nwith the control dataset", yaxt="n")
198 minX <- min(enrich-1)
199 maxX <- max(enrich-1)
200 x = seq(length=11, from=round(minX*10)/10, by=round((maxX-minX)*10)/100)
201 axis(2, at=x,labels=x+1, las=2)
202
203 cat ("Enrichment of genomic regions, ChIP peaks vs Control Peaks:\n")
204 cat (paste(c(names),sep='\t'))
205 cat("\n")
206 cat (paste(c(yChIPTotal/yControlTotal),sep='\t'))
207 cat("\n\n")
208 if (bootstrap>1) {
209 z <- (yChIPTotal-yControlTotal)/sqrt(colVars)
210 pvalues <- 2*pnorm(-abs(z))
211 cat ("Two-side P-values for each genomic category:\n")
212 cat (paste(c(names),sep='\t'))
213 cat("\n")
214 cat (paste(c(yChIPTotal/yControlTotal),sep='\t'))
215 cat("\n\n")
216 }
217 }
218 }
219 if (ifReg) {
220 n.types <- length(levels(chip.data$"6"))
221 myLevels <- levels(chip.data$"6")
222 nlev <- length(names)
223
224 if (ifControl) {
225 cum = matrix( 0, nrow=length(myLevels)+1, ncol=nlev, byrow = TRUE)
226 cumEnrichTotal = matrix( 0, nrow=length(myLevels), ncol=nlev, byrow = TRUE)
227 cumEnrichControl = matrix( 0, nrow=length(myLevels), ncol=nlev, byrow = TRUE)
228 }else {
229 cum = matrix( 0, nrow=length(myLevels), ncol=nlev, byrow = TRUE)
230 cumEnrichTotal = matrix( 0, nrow=length(myLevels), ncol=nlev, byrow = TRUE)
231 }
232 colReg <-NULL
233 for (r in c(1:length(myLevels))) {
234 tt <- which(chip.data$"6"==myLevels[r])
235 subset.data <- (chip.data[tt,])
236 countTotalSubset <- length(unique(subset.data$"1"))
237 tt <- which(subset.data$"16">0)
238 countTotalEnhSubset <- length(unique(subset.data$"1"[tt]))
239 tt <- which(subset.data$"10">0)
240 countTotalPromSubset <- length(unique(subset.data$"1"[tt]))
241 tt <- which(subset.data$"12">0)
242 countTotalImDownSubset <- length(unique(subset.data$"1"[tt]))
243 tt <- which(subset.data$"18">0)
244 countTotalIntraSubset <- length(unique(subset.data$"1"[tt]))
245 tt <- which(subset.data$"20">0)
246 countTotal5kbDSubset <- length(unique(subset.data$"1"[tt]))
247 tt <- which(subset.data$"28">0)
248 countTotal1IntronISubset <- length(unique(subset.data$"1"[tt]))
249 tt <- which(subset.data$"32">0)
250 countTotalExonsISubset <- length(unique(subset.data$"1"[tt]))
251 tt <- which(subset.data$"36">0)
252 countTotalIntronsISubset <- length(unique(subset.data$"1"[tt]))
253 tt <- which(subset.data$"40">0)
254 countTotalJonctionsISubset <- length(unique(subset.data$"1"[tt]))
255 ySubsetTotal <- c (countTotalEnhSubset/countTotalSubset,countTotalPromSubset/countTotalSubset, countTotalImDownSubset/countTotalSubset,countTotalIntraSubset/countTotalSubset,countTotal5kbDSubset/countTotalSubset,countTotal1IntronISubset/countTotalSubset,countTotalExonsISubset/countTotalSubset,countTotalIntronsISubset/countTotalSubset,countTotalJonctionsISubset/countTotalSubset)
256 for (i in c(1:nlev)) {
257 cum[r,i] <- ySubsetTotal[i]
258 cumEnrichTotal[r,i] <- ySubsetTotal[i]/yChIPTotal[i]
259 if (ifControl) {
260 cumEnrichControl[r,i] <- ySubsetTotal[i]/yControlTotal[i]
261 }
262 }
263 colReg[r]<-myColor[r+3]
264 }
265 if (ifControl) {
266 for (i in c(1:nlev)) {
267 cum[4,i] <- yControlTotal[i]
268 }
269 }
270
271 cat ("Proportion of genes with a peak from the ChIP dataset in a given genomic region:\n")
272 for (r in c(1:length(myLevels))) {
273 cat (paste(myLevels[r],":\n",sep=""))
274 cat (paste(c(names),sep='\t'))
275 cat("\n")
276 cat (paste(c(cum[r,] ),sep='\t'))
277 cat("\n")
278 }
279 cat("\n")
280 par(mar=c(5.1, 7.1, 4.1, 2.1))
281 if (ifControl) {
282 barx <- barplot(cum,xlab="",beside=TRUE, col=c(colReg, colors()[334]), legend = c(myLevels,"Control"), names.arg=c(names),ylab="Proportion of genes with a peak")
283 cat ("Proportion of genes with a peak from the Control dataset in a given genomic region:\n")
284 cat (paste(c(names),sep='\t'))
285 cat("\n")
286 cat (paste(c(yControlTotal),sep='\t'))
287 cat("\n\n")
288 if (bootstrap>1) {
289 wiskers <- cum-cum
290 wiskers[nrow(wiskers),] <- sqrt(colVars)*1.96
291 error.bar(barx,cum,wiskers)
292 cat ("Standard deviation for the Control dataset in a given genomic region:\n")
293 cat (paste(c(names),sep='\t'))
294 cat("\n")
295 cat (paste(c(sqrt(colVars)),sep='\t'))
296 cat("\n\n")
297 }
298 } else {
299 barplot(cum,xlab="",beside=TRUE, col=c(colReg), legend = c(myLevels), names.arg=c(names),ylab="Proportion of genes with a peak")
300 }
301 barplot(cumEnrichTotal-1,xlab="",beside=TRUE, col=c(colReg), names.arg=c(names),ylab="Enrichment in comparison\nwith the total gene set", yaxt="n")
302 minX <- min(cumEnrichTotal-1)
303 maxX <- max(cumEnrichTotal-1)
304 x = seq(length=11, from=round(minX*10)/10, by=round((maxX-minX)*10)/100)
305 axis(2, at=x,labels=x+1, las=2)
306
307 cat ("Enrichment of genomic regions, Transcriptional categories vs All Genes:\n")
308 for (r in c(1:length(myLevels))) {
309 cat (paste(myLevels[r],":\n",sep=""))
310 cat (paste(c(names),sep='\t'))
311 cat("\n")
312 cat (paste(c(cumEnrichTotal[r,]),sep='\t'))
313 cat("\n")
314 }
315 cat("\n")
316
317 if (ifControl) {
318 barplot(cumEnrichControl-1,xlab="",beside=TRUE, col=c(colReg), names.arg=c(names),ylab="Enrichment in comparison\nwith control", yaxt="n")
319 minX <- min(cumEnrichControl-1)
320 maxX <- max(cumEnrichControl-1)
321 x = seq(length=11, from=round(minX*10)/10, by=round((maxX-minX)*10)/100)
322 axis(2, at=x,labels=x+1, las=2)
323 cat ("Enrichment of genomic regions, ChIP peaks vs Control Peaks:\n")
324 for (r in c(1:length(myLevels))) {
325 cat (paste(myLevels[r],":\n",sep=""))
326 cat (paste(c(names),sep='\t'))
327 cat("\n")
328 cat (paste(c(cumEnrichControl[r,]),sep='\t'))
329 cat("\n")
330 }
331 cat("\n")
332 if (bootstrap>1) {
333 cat ("Two-side P-values for each genomic category:\n")
334 for (r in c(1:length(myLevels))) {
335 z <- (cum[r,]-yControlTotal)/sqrt(colVars)
336 pvalues <- 2*pnorm(-abs(z))
337 cat (paste(myLevels[r],":\n",sep=""))
338 cat (paste(c(names),sep='\t'))
339 cat("\n")
340 cat (paste(c(pvalues),sep='\t'))
341 cat("\n")
342 }
343 }
344 }
345 }
346 sink() #stop sinking :)
347 dev.off()
348