-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathBCA_GWAS_bmi_interaction_MPI.R
218 lines (175 loc) · 8.21 KB
/
BCA_GWAS_bmi_interaction_MPI.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
############################################
#### genetic risk prediction for BE/EA #####
#### grant application 2020 #####
############################################
#request large node
#salloc -t 2-1 --mem-per-cpu 32G -n 21 --partition=largenode mpirun -n 1 Rscript BCA_GWAS_bmi_interaction_MPI.R
rm(list=ls())
#install.packages("WGCNA") #bigtranspose
#install.packages("/fh/fast/dai_j/CancerGenomics/Tools/Rpackages/plink2R-master/plink2R_1.1.tar.gz", repos = NULL, type = "source")
# library(plink2R)
# dat_bca=read_plink("/fh/fast/dai_j/BEACON/BeagessCambridgeAmos/bca_filteredMAF_20Feb2015",impute="avg")
# bim=dat_bca$bim
# genotype=dat_bca$bed
# fam=dat_bca$fam
# #rm(dat_bca)
# idx=bim$V1 %in% 1:23 #23:X
# bim=bim[idx,]
# genotype=genotype[,idx]
# allgenotype=genotype
# save(allgenotype,fam,bim,file="../result/bca_filteredMAF_20Feb2015.RData",version=2)
# library("WGCNA")
# tmp=transposeBigData(allgenotype, blocksize = 20000)
# all(as.numeric(tmp[2500,])==allgenotype[,2500])
# allgenotype=as.data.frame(tmp)
# save(allgenotype,fam,bim,file="../result/bca_filteredMAF_20Feb2015_t.RData")
load("/fh/fast/dai_j/BEACON/BEACON_GRANT/result/bca_filteredMAF_20Feb2015_t.RData")
library(readxl)
sampletable=read_excel("/fh/fast/dai_j/BEACON/BEACON_GRANT/data/PLINKinputCombo_bca_07Feb2018.xls",1)
sampletable <- data.frame(sampletable)
for (i in 1:ncol(sampletable)) sampletable[which(sampletable[,i]==-9),i]=NA
allsamples=intersect(fam$V2,sampletable$localid)
idx=match(allsamples,sampletable$localid)
sampletable=sampletable[idx,]
rownames(sampletable)=sampletable$localid
Covariate=sampletable[,colnames(sampletable) %in% c("sex","age","ev1_bca","ev2_bca","ev3_bca","ev4_bca","bmi_recent_healthy")]
colnames(Covariate[which(colnames(Covariate)=="ev1_bca")])="pc1"
colnames(Covariate[which(colnames(Covariate)=="ev2_bca")])="pc2"
colnames(Covariate[which(colnames(Covariate)=="ev3_bca")])="pc3"
colnames(Covariate[which(colnames(Covariate)=="ev4_bca")])="pc4"
rownames(Covariate)=sampletable$localid
Covariate$age[is.na(Covariate$age)]=mean(Covariate$age,na.rm = T)
nprobes <- nrow(allgenotype)
sink("../result/MPI_output.txt") #changed-
### Prepare Rmpi environment, DO NOT CHANGE ###
Sys.getenv(c("SLURM_SUBMIT_DIR"))
Sys.getenv(c("HOST", "SLURM_JOB_ID", "SLURM_NODELIST","SLURM_NNODES", "SLURM_NTASKS",
"SLUR M_CPUS_PER_TASK","SLURM_CPUS_ON_NODE","SLURM_NTASKS_PER_NODE",
"SLURM_TASK_PID", "SLURM_ PARTITION"))
#
library('Rmpi')
njob <- mpi.universe.size() - 1L
mpi.spawn.Rslaves(nslaves=njob)
.Last <- function(){
if (is.loaded("mpi_initialize")){
if (mpi.comm.size(1) > 0){
print("Please use mpi.close.Rslaves() to close slaves.")
mpi.close.Rslaves()
}
print("Please use mpi.quit() to quit R")
.Call("mpi_finalize")
}
}
#
starttime<-Sys.time();starttime
compute_pvalue=function(i=1,pheno,genodat,covdat)
{
res=data.frame(p1=NA,beta1=NA,p2=NA,beta2=NA,p3=NA)
genotype=as.numeric(genodat[i,])
#covdat$cig_smk_ever <- 1-covdat$cig_smk_ever
GE <- genotype*covdat$bmi_recent_healthy
fit1 <- glm(I(pheno==2)~GE+genotype+bmi_recent_healthy+sex+age+ev1_bca+ev2_bca+ev3_bca+ev4_bca,family=binomial,data=covdat)
fit2 <- glm(I(pheno==2)~bmi_recent_healthy+sex+age+ev1_bca+ev2_bca+ev3_bca+ev4_bca,family=binomial,data=covdat)
res$p3 <- pchisq(fit2$deviance-fit1$deviance,df=2,lower.tail = F)
if ("GE" %in% rownames(summary(fit1)$coef))
{
res$p1 <- summary(fit1)$coef[2,4]
res$beta1 <- summary(fit1)$coef[2,1]
res$p2 <- summary(fit1)$coef[3,4]
res$beta2 <- summary(fit1)$coef[3,1]
}
return(res)
}
mpi.bcast.Robj2slave(sampletable)
mpi.bcast.Robj2slave(Covariate)
mpi.bcast.Robj2slave(compute_pvalue)
mpi_compute_pvalue <- function(nsnps=1000){
nchunks=ceiling(nprobes/nsnps)
print(paste0("number of total:",nchunks))
rows=1:nprobes
out=NULL
for (ii in 1:nchunks)
{
if (ii%% 50==0)cat(ii,"..")
if (ii<nchunks)
{
seqs=rows[((ii-1)*nsnps+1):(ii*nsnps)]
}else
{
seqs=rows[((ii-1)*nsnps+1):length(rows)]
}
out1 <- matrix(NA,length(seqs),20)
rownames(out1)=rownames(allgenotype)[seqs]
Covariate1 <- Covariate
pheno=sampletable$phenoBE_bca
genodat=allgenotype[seqs,]
idx=!is.na(pheno) & !is.na(Covariate1$bmi_recent_healthy)
pheno=pheno[idx]
genodat=genodat[,idx]
covdat=Covariate1[idx,]
tmp=mpi.parSapply(X=1:length(seqs),FUN=compute_pvalue,pheno=pheno,genodat=genodat,covdat=covdat,job.num=njob)
out1[,1:5]=matrix(unlist(tmp),ncol=5,byrow = T)
pheno=sampletable$phenoEA_bca
genodat=allgenotype[seqs,]
idx=!is.na(pheno)& !is.na(Covariate1$bmi_recent_healthy)
pheno=pheno[idx]
genodat=genodat[,idx]
covdat=Covariate1[idx,]
tmp=mpi.parSapply(X=1:length(seqs),FUN=compute_pvalue,pheno=pheno,genodat=genodat,covdat=covdat,job.num=njob)
out1[,6:10]=matrix(unlist(tmp),ncol=5,byrow = T)
pheno=sampletable$phenoEABE_bca
genodat=allgenotype[seqs,]
idx=!is.na(pheno)& !is.na(Covariate1$bmi_recent_healthy)
pheno=pheno[idx]
genodat=genodat[,idx]
covdat=Covariate1[idx,]
tmp=mpi.parSapply(X=1:length(seqs),FUN=compute_pvalue,pheno=pheno,genodat=genodat,covdat=covdat,job.num=njob)
out1[,11:15]=matrix(unlist(tmp),ncol=5,byrow = T)
pheno=sampletable$phenoEA_bca
pheno[is.na(pheno)] = 3
pheno[pheno==1] = NA
genodat=allgenotype[seqs,]
idx=!is.na(pheno)& !is.na(Covariate1$bmi_recent_healthy)
pheno=pheno[idx]
genodat=genodat[,idx]
covdat=Covariate1[idx,]
tmp=mpi.parSapply(X=1:length(seqs),FUN=compute_pvalue,pheno=pheno,genodat=genodat,covdat=covdat,job.num=njob)
out1[,16:20]=matrix(unlist(tmp),ncol=5,byrow = T)
out=rbind(out,out1)
}
return(out)
}
Sys.time()
BCAout=mpi_compute_pvalue()
sink()
save(BCAout,file="/fh/fast/dai_j/BEACON/BEACON_GRANT/result/BCA_gwas_bmi_interaction_output.RData")
## quit program
mpi.close.Rslaves()
mpi.quit()
quit()
#rm(list=ls())
#load("/fh/fast/dai_j/BEACON/BEACON_GRANT/result/BCA_gwas_bmi_interaction_output.RData")
#mpoint <- ceiling(quantile(1:nrow(BCAout),0.95)/200)
#pindex <- c(1,(1:mpoint)*200,(mpoint*200+1):nrow(BCAout))
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,1],decreasing=T),1],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,3],decreasing=T),3],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,5],decreasing=T),5],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,6],decreasing=T),6],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,8],decreasing=T),8],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,10],decreasing=T),10],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,11],decreasing=T),11],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,13],decreasing=T),13],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,15],decreasing=T),15],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,20],decreasing=T),20],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)
#plot(-log((nrow(BCAout):1)/nrow(BCAout),base=10)[pindex],-log(BCAout[order(BCAout[,7],decreasing=T),7],base=10)[pindex],xlab="-log10 expected p-value",ylab="-log10 observed p-value")
#abline(0,1)