-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathgenerate_summary_statistics.R
executable file
·236 lines (185 loc) · 11.6 KB
/
generate_summary_statistics.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
#!/usr/bin/env Rscript
# generate_summary_statistics.R
# Generate dxy and partitioning statistics for models generated using run_model_combinations.py
# Written for "Evaluating the use of ABBA-BABA statistics to locate introgressed loci"
# by Simon H. Martin, John W. Davey and Chris D. Jiggins
# Simon Martin: [email protected]
# John Davey: [email protected]
# November-December 2013, May 2014, August 2014
library(parallel)
suppressMessages(library(reshape))
library(plyr)
library(optparse)
library(stringr)
library(tools)
options<-list(make_option(c("-t","--threads"),default=1,help="Number of threads [default %default]",metavar="numthreads"),
make_option(c("-m","--modelfiles"),default="",type="character",help="Folder of CSV files with model output"),
make_option(c("-r","--realdata"),default="",type="character",help="CSV file containing windows from real genome"),
make_option(c("-l","--listofmodels"),default="",type="character",help="CSV file of model parameter combinations")
)
opt<-parse_args(OptionParser(option_list=options))
if (opt$listofmodels == "" & opt$realdata == "") stop("Please specify a list of model parameter combinations in CSV format with -l")
if (opt$modelfiles == "" & opt$realdata == "") stop("Please specify a folder of model files with -m or a CSV file containing real data with -r")
if (opt$modelfiles != "" & !file.exists(opt$modelfiles)) stop(paste("Directory", opt$modelfiles, "does not exist! Please specify a valid directory with -m"))
if (opt$realdata != "" & !file.exists(opt$realdata)) stop(paste("Real data file", opt$realdata, "does not exist! Please specify a valid file with -r"))
partition.x<-function(model,stat,threshold=0.1,win) {
outnum<-threshold*win
Partition<-sapply(rank(-model[,stat],ties.method="random"), function(x) {if (x<=outnum) "_Outlier" else "_Background"})
Partition<-paste0(stat,Partition)
model<-cbind(model,Partition)
names(model)[names(model)=="Partition"]<-paste0(stat,"_Partition")
model
}
run.tests<-function(values,groups,test.name,test.func) {
# Assumes order of groups is Background, Outlier
formula<-values~groups
two.tailed<-test.func(formula,alternative="two.sided",conf.int=TRUE)
if (length(two.tailed$estimate)==1) mean.diff<-two.tailed$estimate else mean.diff<-two.tailed$estimate[1] - two.tailed$estimate[2]
out<-data.frame(
mean.diff=mean.diff,
p.two.tailed=two.tailed$p.value,
p.background.higher=test.func(formula,alternative="greater")$p.value,
p.background.lower=test.func(formula,alternative="less")$p.value
)
names(out)<-paste(test.name,names(out),sep=".")
out
}
summarise.tests<-function(model.sum,model.stats,model.partition,test.name,test.func) {
model.tests<-t(apply(model.stats,2,function(x) unlist(run.tests(x,model.partition,test.name,test.func))))
model.tests<-cbind(Variable=rownames(model.tests),model.tests)
merge(model.sum,model.tests,by="Variable",sort=FALSE)
}
get.model.summary<-function(models,column,pars=NULL) {
model.stats<-models[c("P1P2_dxy","P1P3_dxy","P2P3_dxy")]
model.partition<-models[,column]
model.means<-melt(aggregate(model.stats,list(model.partition),mean,na.rm=TRUE),id="Group.1")
model.sds<-melt(aggregate(model.stats,list(model.partition),sd,na.rm=TRUE),id="Group.1")
model.ses<-melt(aggregate(model.stats,list(model.partition), function(x) {x<-na.omit(x);sqrt(var(x)/length(x))}),id="Group.1")
model.sum<-cbind(pars,model.means,model.sds$value,model.ses$value)
names(model.sum)<-c(names(pars),"Model","Variable","Mean","SD","SE")
model.sum<-summarise.tests(model.sum,model.stats,model.partition,"t",t.test)
model.sum<-summarise.tests(model.sum,model.stats,model.partition,"w",wilcox.test)
model.sum
}
calc.fD0<-function(df, fstat) {
apply(df,1,function(x) if ((is.na(x["D"])) || (as.numeric(x["D"])<0)) NA else as.numeric(x[fstat]))
}
get.fs<-function(models, altprop, windows) {
models$fGD0<-calc.fD0(models,"fG")
models$fhomD0<-calc.fD0(models,"fhom")
models$fdD0<-calc.fD0(models,"fd")
models<-partition.x(models,"D",altprop,windows)
models<-partition.x(models,"fGD0",altprop,windows)
models<-partition.x(models,"fhomD0",altprop,windows)
models<-partition.x(models,"fdD0",altprop,windows)
models
}
mean.pc<-function(stat) {
mean(stat, na.rm=TRUE) * 100
}
sd.pc<-function(stat) {
sd(stat, na.rm=TRUE) * 100
}
summarise.model<-function(model, modeldir, seqs) {
filestub = paste0("Alternate_t123-", model$Alternate_t123, "_t23-", model$Alternate_t23, ".Background_t123-", model$Background_t123, "_t21-", model$Background_t21, ".*.", seqs, ".csv")
filename = dir(modeldir, filestub, full.names=TRUE)
pars<-data.frame(str_match(filename,"Alternate_t123-(.+)_t23-(.+)\\.Background_t123-(.+)_t21-(.+)\\.(.+)\\..+\\.csv"),stringsAsFactors=FALSE)
names(pars)<-c("File","Alternate_t123","Alternate_t23","Background_t123","Background_t21","Windows")
pars$Windows<-as.numeric(pars$Windows)
models<-read.csv(filename)
models$Model<-factor(str_match(models$Model,"(.+)_t123")[,2],levels=c("Background","Alternate"))
models$Model<-revalue(models$Model,c(Background="Real_Background",Alternate="Real_Outlier"))
models<-get.fs(models, 0.1, pars$Windows)
model.summary<-get.model.summary(models,"Model",pars)
D.summary<-get.model.summary(models,"D_Partition",pars)
fGD0.summary<-get.model.summary(models,"fGD0_Partition",pars)
fhomD0.summary<-get.model.summary(models,"fhomD0_Partition",pars)
fdD0.summary<-get.model.summary(models,"fdD0_Partition",pars)
model.sum<-rbind(model.summary, D.summary, fGD0.summary, fhomD0.summary, fdD0.summary)
is.D.positive<-models$D>=0
real.outliers<-models$Model=="Real_Outlier"
model.comp<-data.frame(pars,
Stat=c("D","fGD0","fhomD0","fdD0"),
Stat.Mean=c(mean.pc(models$D), mean.pc(models$fGD0), mean.pc(models$fhomD0), mean.pc(models$fdD0)),
Stat.SD=c(sd.pc(models$D), sd.pc(models$fGD0), sd.pc(models$fhomD0), sd.pc(models$fdD0)),
Dpos=sum(is.D.positive,na.rm=TRUE),
DposAlternate=sum(is.D.positive & real.outliers,na.rm=TRUE),
DposOutlier=c(sum(is.D.positive & models$D_Partition=="D_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fGD0_Partition=="fGD0_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fhomD0_Partition=="fhomD0_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fdD0_Partition=="fdD0_Outlier", na.rm=TRUE)),
OutlierAlternate=c(sum(real.outliers & models$D_Partition=="D_Outlier",na.rm=TRUE),
sum(real.outliers & models$fGD0_Partition=="fGD0_Outlier",na.rm=TRUE),
sum(real.outliers & models$fhomD0_Partition=="fhomD0_Outlier",na.rm=TRUE),
sum(real.outliers & models$fdD0_Partition=="fdD0_Outlier",na.rm=TRUE))
)
model.comp$OutlierAlternatePC<-model.comp$OutlierAlternate/model.comp$DposOutlier*100
list(model.sum,model.comp)
}
summarise.null<-function(model, modeldir, seqs) {
filepattern = paste0("^Background_t123-", model$Background_t123, "_t21-", model$Background_t21,".*.", seqs, ".csv")
filename = dir(modeldir, filepattern, full.names=TRUE)
pars<-data.frame(str_match(filename,"Background_t123-(.+)_t21-(.+)\\.(.+)\\..+\\.csv"),stringsAsFactors=FALSE)
names(pars)<-c("File","Background_t123","Background_t21","Windows")
pars$Windows<-as.numeric(pars$Windows)
models<-read.csv(filename)
models<-get.fs(models, 0.1, pars$Windows)
D.summary<-get.model.summary(models,"D_Partition",pars)
fGD0.summary<-get.model.summary(models,"fGD0_Partition",pars)
fhomD0.summary<-get.model.summary(models,"fhomD0_Partition",pars)
fdD0.summary<-get.model.summary(models,"fdD0_Partition",pars)
model.sum<-rbind(D.summary, fGD0.summary, fhomD0.summary, fdD0.summary)
is.D.positive<-models$D>=0
model.comp<-data.frame(pars,
Stat=c("D","fGD0","fhomD0","fdD0"),
Stat.Mean=c(mean.pc(models$D), mean.pc(models$fGD0), mean.pc(models$fhomD0), mean.pc(models$fdD0)),
Stat.SD=c(sd.pc(models$D), sd.pc(models$fGD0), sd.pc(models$fhomD0), sd.pc(models$fdD0)),
Dpos=sum(is.D.positive,na.rm=TRUE),
DposOutlier=c(sum(is.D.positive & models$D_Partition=="D_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fGD0_Partition=="fGD0_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fhomD0_Partition=="fhomD0_Outlier", na.rm=TRUE),
sum(is.D.positive & models$fdD0_Partition=="fdD0_Outlier", na.rm=TRUE))
)
list(model.sum,model.comp)
}
process.alt<-function(modeldir, model.df, seqs) {
model.list<-split(model.df, 1:nrow(model.df)) # Process df by rows
csv.summary<-mclapply(model.list, summarise.model, modeldir=modeldir, seqs=seqs, mc.cores=opt$threads)
models<-rbind.fill(lapply(csv.summary,function(x) x[[1]]))
model.comp<-rbind.fill(lapply(csv.summary,function(x) x[[2]]))
models<-merge(models,unique(model.df[,c("Background_t123","Background_t21","Alternate_t123","Alternate_t23","ModelType")]),by=c("Background_t123","Background_t21","Alternate_t123","Alternate_t23"))
write.table(models,file=paste0(opt$modelfiles,".alternate_models.dxy.summary.", seqs, ".tsv"),sep="\t",row.names=FALSE)
model.comp<-merge(model.comp,unique(model.df[,c("Background_t123","Background_t21","Alternate_t123","Alternate_t23","ModelType")]),by=c("Background_t123","Background_t21","Alternate_t123","Alternate_t23"))
write.table(model.comp,file=paste0(opt$modelfiles,".alternate_models.partition.summary.", seqs, ".tsv"),sep="\t",row.names=FALSE)
}
process.null<-function(modeldir, model.df, seqs) {
model.list<-split(model.df, 1:nrow(model.df)) # Process df by rows
csv.summary<-mclapply(model.list, summarise.null, modeldir=modeldir, seqs=seqs, mc.cores=opt$threads)
models<-rbind.fill(lapply(csv.summary,function(x) x[[1]]))
model.comp<-rbind.fill(lapply(csv.summary,function(x) x[[2]]))
models<-merge(models,unique(model.df[,c("Background_t123","Background_t21","ModelType")]),by=c("Background_t123","Background_t21"))
write.table(models,file=paste0(opt$modelfiles,".null_models.dxy.summary.", seqs, ".tsv"),sep="\t",row.names=FALSE)
model.comp<-merge(model.comp,unique(model.df[,c("Background_t123","Background_t21","ModelType")]),by=c("Background_t123","Background_t21"))
write.table(model.comp,file=paste0(opt$modelfiles,".null_models.partition.summary.", seqs, ".tsv"),sep="\t",row.names=FALSE)
}
options(width=200)
if (opt$modelfiles != "") {
model.list<-read.csv(opt$listofmodels,stringsAsFactors=FALSE,colClasses="character")
for (seqs in c("ms","sg")) {
process.alt(opt$modelfiles, model.list[model.list$ModelType != "Null model",], seqs)
process.null(opt$modelfiles, model.list[model.list$ModelType == "Null model",], seqs)
}
}
if (opt$realdata != "") {
real<-read.csv(opt$realdata)
real$fdD0<-calc.fD0(real, "fd")
real<-partition.x(real,"D",0.1,nrow(real))
real<-partition.x(real,"fGD0",0.1,nrow(real))
real<-partition.x(real,"fhomD0",0.1,nrow(real))
real<-partition.x(real,"fdD0",0.1,nrow(real))
pars<-data.frame(File=opt$realdata,Windows=nrow(real))
real.stats<-rbind(get.model.summary(real,"D_Partition",pars),get.model.summary(real,"fGD0_Partition",pars),get.model.summary(real,"fhomD0_Partition",pars),get.model.summary(real,"fdD0_Partition",pars))
write.table(real.stats,file=paste0(file_path_sans_ext(basename(opt$realdata)),".dxy.summary.tsv"),sep="\t",row.names=FALSE)
}
warninglist<-warnings()
if(!(is.null(warninglist))) warninglist