-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathGBS_binning.r
317 lines (235 loc) · 11.7 KB
/
GBS_binning.r
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
data_name="" #your data name
data.total<-read.table(data_name,header=FALSE,sep="\t")
data.total=as.matrix(data.total)
ids<-data.total[,1]
chromosome=ids
chromosome=strsplit(as.character(chromosome),":")
distance.total=sapply(chromosome,function(x) x[2])
chromosome=sapply(chromosome,function(x) x[1])
distance.total=as.numeric(as.character(distance.total))
chromosome=as.numeric(as.character(chromosome))
data.total=(data.total[,-1])
data.total[which(data.total=="A")]=1
data.total[which(data.total=="B")]=0
data.total[which(data.total=="H")]=0.5
data.total[which(data.total=="-")]=NA
class(data.total) <- "numeric"
break_points=matrix(list(),8)
for(i in 1:8){break_points[[i]]=0}
bin_starts=integer()
NofislandsR_total=integer()
binC=1
#x <- scan(break_points_file_name, what="", sep="\n")
#pre_break_points <- strsplit(x, ",[[:space:]]+")
#pre_joinmap_choice=read.table("joinmap_choice",header=FALSE)
#pre_joinmap_choice=as.numeric(as.character(as.matrix(pre_joinmap_choice)))
#pdf(paste("pdf_file_name.pdf",sep=""))
for(chr in min(chromosome):max(chromosome)){
print(paste("chromosome",chr))
data=data.total[which(chromosome==chr),]
data=as.matrix(data)
mode(data)<-"numeric"
distance=distance.total[which(chromosome==chr)]
mode(distance)<-"numeric"
ids.chr=ids[which(chromosome==chr)]
aftersmoothingA=matrix(0,length(data[1,]),length(data[,1])) #this will be the final result
lastsnp=length(data[,1]) #number of snps
Nofislands=array(0,dim=c(binC,length(data[1,])))#for the distribution of the number of contigous regions
minD=50000
maxD=5000000
Dseq=10000000*1/5
#block when using a single value# Dseq=seq.int(minD,maxD,length.out=binC) #a sequence of bin sizes
distanceIndex=array(0,dim=c(length(data[,1]),binC,2))
for(x in 1:length(data[,1])){ #this loop is for assigning proper flanking snps according to the given bin size
k=length(Dseq)
temp=intersect(which(distance[x]-distance<=Dseq[k]),which(distance[x]-distance>=0))
if(length(temp)==0){begin=distanceIndex[x-1,k,1]
}else{ begin=min(temp)
}
temp=intersect(which(distance-distance[x]<=Dseq[k]),which(distance-distance[x]>=0))
if(length(temp)==0){end=distanceIndex[x-1,k,2]
}else{ end=max(temp)
}
distanceIndex[x,k,]=c(begin,end)
k=k-1
while(k>0){
#begin=ifelse(begin+1>x,x,begin+1) missing the case that begin stays
distanceIndex[x,k,1]=begin
while(distance[x]-distance[begin]>Dseq[k]){
begin=ifelse(begin+1>x,x,begin+1)
}
k=k-1
}
k=length(Dseq)-1
while(k>0){
distanceIndex[x,k,2]=end
while(distance[end]-distance[x]>Dseq[k]){
end=ifelse(end-1<x,x,end-1)
}
k=k-1
}
}
##the main part
for(j in 1:length(data[1,])){#for each individual
{
# print(paste("at",j))
for(t in 1:binC){# distanceIndex's number of testing threshold
# print(paste("index",t))
newdata=data
#deal with the first region and the last region(to prevent the case that the first bin and the last bin are NAs)
newdata[distanceIndex[1,t,1]:distanceIndex[1,t,2],j]=mean(data[distanceIndex[1,t,1]:distanceIndex[1,t,2],j],na.rm=TRUE)
newdata[distanceIndex[lastsnp,t,1]:distanceIndex[lastsnp,t,2],j]=mean(data[distanceIndex[lastsnp,t,1]:distanceIndex[lastsnp,t,2],j],na.rm=TRUE)
difference=matrix(NA,1,length(data[,1]))#difference of the left and right
for( i in 1:(length(data[,1])-1)){
#compare right and left
difference[i]=mean(newdata[distanceIndex[i,t,1]:i,j],na.rm=TRUE)-mean(newdata[(i):distanceIndex[i,t,2],j],na.rm=TRUE)#changed from i+1 to i in the second term
}
# print(paste("index sub1",t))
#pick peaks
tempD=abs(difference)
peaks=which(tempD==max(tempD,na.rm=TRUE))[1]
peakV=max(tempD,na.rm=TRUE)[1]
if(peakV<0.3){peaks=NA #if it is less than 0.3, it's regarded as small
}else{
tempD[distanceIndex[peaks,t,1]:distanceIndex[peaks,t,2]]=0
}
while(peakV>=0.3){# while having meaningful peaks
peakstemp=which(tempD==max(tempD,na.rm=TRUE))[1]
peakV=max(tempD,na.rm=TRUE)[1]
if(peakV>=0.3){
peaks=c(peaks,peakstemp)
tempD[distanceIndex[peakstemp,t,1]:distanceIndex[peakstemp,t,2]]=0
}
}
# print(paste("index sub2",t))
#smoothing
aftersmoothing=data[,j]
#if(!is.na(peaks)&length(data[,1])-peaks[1]<300){peaks=NA} #short one peak?
if(is.na(peaks)){#if there is no peak
# print(paste("index sub3_1",t))
meantemp=mean(newdata[,j],na.rm=TRUE)
if(is.na(meantemp)){ next }#if the is no data
dis=c(abs(meantemp),abs(meantemp-1),abs(meantemp-0.5))
resultvalue=integer(0)
if(dis[1]<=dis[2]){resultvalue=0
if(dis[3]<=dis[1]){resultvalue=0.5}
}else if(dis[2]<=dis[3]){resultvalue=1
}else{resultvalue=0.5}
# print(paste("index sub3",t))
aftersmoothing=rep(resultvalue,length(aftersmoothing))
}else{#if there are peaks
# print(paste("index sub4",t))
peaks=sort(peaks,decreasing=FALSE)
if(length(peaks)>=2){
l=length(peaks)
for(s in 2:l){
if(distance[peaks[s]]-distance[peaks[s-1]]<(Dseq[t]*2)&abs(mean(newdata[ifelse((s-2)>=1,peaks[(s-2)],1):peaks[s-1],j],na.rm=TRUE)-mean(newdata[peaks[s]:ifelse(s+1<=length(peaks),peaks[s+1],length(data[,1])),j],na.rm=TRUE))<=0.3){ peaks=peaks[-s]} #remove short region & flanking regions with small difference
if((s+1)>length(peaks)){break}
}
}
meantemp=mean(newdata[1:peaks[1],j],na.rm=TRUE)
dis=c(abs(meantemp),abs(meantemp-1),abs(meantemp-0.5))
resultvalue=integer(0)
if(dis[1]<=dis[2]){resultvalue=0
if(dis[3]<=dis[1]){resultvalue=0.5}
}else if(dis[2]<=dis[3]){resultvalue=1
}else{resultvalue=0.5}
aftersmoothing[1:peaks[1]]=resultvalue #the first bin
if(length(peaks)>=2){
for(s in 2:length(peaks)){#form the second bin
meantemp=mean(newdata[peaks[s-1]:peaks[s],j],na.rm=TRUE)
dis=c(abs(meantemp),abs(meantemp-1),abs(meantemp-0.5))
resultvalue=integer(0)
if(dis[1]<=dis[2]){resultvalue=0
if(dis[3]<=dis[1]){resultvalue=0.5}
}else if(dis[2]<=dis[3]){resultvalue=1
}else{resultvalue=0.5}
aftersmoothing[(peaks[s-1]+1):peaks[s]]=resultvalue
}
}
meantemp=mean(newdata[peaks[length(peaks)]:length(data[,1]),j],na.rm=TRUE)
dis=c(abs(meantemp),abs(meantemp-1),abs(meantemp-0.5))
resultvalue=integer(0)
if(dis[1]<=dis[2]){resultvalue=0
if(dis[3]<=dis[1]){resultvalue=0.5}
}else if(dis[2]<=dis[3]){resultvalue=1
}else{resultvalue=0.5}
## print(paste("index sub5",t))
aftersmoothing[(peaks[length(peaks)]+1):length(data[,1])]=resultvalue
}#not NA
# par(mfrow=c(2,1))
# print(paste("index sub6",t))
aftersmoothingA[j,]=aftersmoothing
# plot(aftersmoothing,ylim=c(0,1))
# plot(data[,j],col="green",xlab="", ylab="",lwd=1)
#count the number of islands
bin_starts_temp=integer()
for(q in 2:lastsnp){
if(aftersmoothing[q-1]!=aftersmoothing[q]){ Nofislands[t,j]=Nofislands[t,j]+1 ;
bin_starts_temp=c(bin_starts_temp,distance[q])
break_points[[chr]]=c(break_points[[chr]],distance[q])
}
#
}
bin_starts=rbind(bin_starts,c(chr,j,t,bin_starts_temp))
# print(paste("index sub7",t))
if(chr==0){#closing this section because there are too many plots
plot(as.numeric(as.matrix(aftersmoothingA[j,])),col="green",pch="*",ylim=c(0,1),main=paste("binning index:",t,"chromsome:",chr,"individual:",j));
#abline(v=which(distance%in%as.numeric(pre_break_points[[chr]])),col="grey")
#abline(v=pre_joinmap_choice,col="grey")
}
# if(length(bin_starts_temp)>0){ plot(as.numeric(as.matrix(aftersmoothingA[j,])),col="green",pch="*",ylim=c(0,1),main=paste("binning index:",t,"chromsome:",chr,"individual:",j));abline(v=which(distance%in%as.numeric(pre_break_points[[chr]])),col="grey")
# }
abc=1
}#t iteration:distanceIndex's number of testing threshold
if(chr==0){
plot(as.numeric(as.matrix(data[,j])),col="blue",pch="*",ylim=c(0,1),main=paste("binning index:",t,"chromsome:",chr,"individual:",j)); #plots the original data
#abline(v=which(distance%in%as.numeric(pre_break_points[[chr]])),col="grey")
#abline(v=pre_joinmap_choice,col="grey")
}
}#j iteration/ individual
}
#dev.off()
#for the bin distributoin
NofislandsR=apply(Nofislands,1,sum) #to check the number of contiguous regions based on each bin size, plot this
NofislandsR_total=cbind(NofislandsR_total,NofislandsR)
##plot(NofislandsR,main=chr)
#below should be applied applying the optimal t
break_points[[chr]]=sort(unique(break_points[[chr]]))
break_points[[chr]]= break_points[[chr]][-1]
bp_start= 1
bp_end= which(distance==break_points[[chr]][1])
middle=round((bp_start+bp_end)/2)
joinmap_choice=middle
for(bps in 2:length(break_points[[chr]])){
bp_start= which(distance==break_points[[chr]][bps-1])
bp_end= which(distance==break_points[[chr]][bps])
middle=round((bp_start+bp_end)/2)
joinmap_choice=c(joinmap_choice,middle)
}
bp_start= which(distance==break_points[[chr]][length(break_points[[chr]])])
bp_end= length(break_points[[chr]])
middle=round((bp_start+bp_end)/2)
joinmap_choice=c(joinmap_choice,middle)
joinmap_choice=sort(unique(joinmap_choice))
break_points.smoothed=t(aftersmoothingA[,joinmap_choice])
break_points.data=data[joinmap_choice,]
break_points.ids=ids.chr[joinmap_choice]
mixed=integer()
for(s in 1:length(joinmap_choice)){
mixed=cbind(mixed,break_points.smoothed[s,],break_points.data[s,])
colnames(mixed)[(2*(s-1)+1):(2*s)]=c(ids.chr[s],ids.chr[s])
}
break_points.smoothed[which(break_points.smoothed==1)]="A"
break_points.smoothed[which(break_points.smoothed==0)]="B"
break_points.smoothed[which(break_points.smoothed==0.5)]="H"
break_points.smoothed[which(is.na(break_points.smoothed)==TRUE)]="-"
if(chr==0){write.table((joinmap_choice),file="joinmap_choice",quote=FALSE,row.names=FALSE,col.names=FALSE)}
write.table(mixed,file=paste(data_name,"chosen_smoothed_original_comparison.",chr,sep=""),col.names=colnames(mixed),row.names=FALSE,quote=FALSE)
write.table(paste("NAME=Cr\nPOPT=F2\nNLOC=",length(break_points.smoothed[,1]),"\nNIND=",length(break_points.smoothed[1,]),sep=""),file=paste(data_name,".chosen_smoothed.chr",chr,sep=""),row.names= FALSE,col.names=FALSE,quote=FALSE)
write.table((break_points.smoothed),file=paste(data_name,".chosen_smoothed.chr",chr,sep=""),row.names= break_points.ids,col.names=FALSE,quote=FALSE,append=TRUE,sep="\t")
#write.table(break_points.data,file=paste(data_name,".chosen_smoothed.chr",chr,sep=""),col.names= break_points.ids,row.names=FALSE,quote=FALSE)
write.table(break_points.ids,file=paste(data_name,".chosen_markers",sep=""),row.names=FALSE,col.names=FALSE,quote=FALSE,append=TRUE)
}#chromsome
NofislandsR_total
#dev.off()