forked from ramachandran-lab/ESNN
-
Notifications
You must be signed in to change notification settings - Fork 0
/
simu_example.R
108 lines (91 loc) · 3.09 KB
/
simu_example.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
rm(list = ls())
binary_simu = function(ind, obs, nsnp, pve, ncausal_features, k, min_maf){
#simulate X
maf <- min_maf + (0.5-min_maf)*runif(nsnp)
Geno <- (runif(ind*nsnp) < maf) + (runif(ind*nsnp) < maf)
Geno <- matrix(as.double(Geno),ind,nsnp,byrow = TRUE)
#sample causal
causal_snps = sample(c(1:nsnp), ncausal_features)
Xeff = Geno[, causal_snps]
#scale according to MAFs
ratio = rep(0, ncausal_features)
for (i in 1:ncausal_features) {
temp_Xeff = Xeff[,i]
ratio[i] = 2*length(which(temp_Xeff == 2))/length(which(temp_Xeff == 1))
}
z = rep(0, ind)
b2 = rnorm(ncausal_features)
b1 = -1*b2*ratio
for (i in 1:ncausal_features) {
temp_Xeff = Xeff[,i]
idx = which(temp_Xeff == 1)
z[idx] = z[idx] + b1[i]
idx = which(temp_Xeff == 2)
z[idx] = z[idx] + b2[i]
}
z = z*sqrt(pve)/sd(z) + rnorm(ind, 0, sqrt(1-pve))
z = (z - mean(z))/sd(z)
### Set the Threshold ###
thresh=qnorm(1-k,mean=0,sd=1)
### Find the Number of Cases and Controls ###
n.cases = sum(z>thresh); n.cases/length(z)
n.controls = sum(z<=thresh); n.controls/length(z)
### Subsample a particular number of cases and controls ###
print(length(which(z>thresh)))
cases = sample(which(z>thresh),obs,replace = FALSE)
controls = sample(which(z<=thresh),obs,replace = FALSE)
y = c(rep(1, obs), rep(0, obs))
X = Geno[c(cases,controls),]
for (i in 1:nsnp) {
X[,i] = (X[,i] - mean(X[,i]))/sd(X[,i])
}
rm(Geno, Xeff)
return(list(X, y, causal_snps))
}
ind = 1e6
obs = 2500
nsnp = 200
ncausal_features = 5
pve = 0.4
k = 0.1
min_maf = 0.05
results = binary_simu(ind, obs, nsnp, pve, ncausal_features, k, min_maf)
X = results[[1]]
Y = results[[2]]
causal_snps = results[[3]]
#
write.table(X, file = 'data/X_binary.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
write.table(Y, file = 'data/Y_binary.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
write.table(causal_snps, file = 'data/causal_binary.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
regression_simu = function(num_example, num_feature, pve, ncausal_features){
X = matrix(0, nrow = num_example, ncol = num_feature)
for (i in 1:num_feature) {
X[,i] = rnorm(num_example, 0, 1)
}
causal_feature = sample(c(1:num_feature), ncausal_features)
Xeff = X[,causal_feature]
Yeff = Xeff[,1] + cos(Xeff[,1])+Xeff[,2]*Xeff[,3] + sin(Xeff[,4] + Xeff[,5])
Y = Yeff*sqrt(pve)/sd(Yeff) + rnorm(num_example, 0, sqrt(1-pve))
Y = (Y - mean(Y))/sd(Y)
return(list(X, Y, causal_feature))
}
corr = 0.2
results = regression_simu(5000, 100, 0.6, 5)
X = results[[1]]
Y = results[[2]]
causal_snps = results[[3]]
#
write.table(X, file = 'data/X_reg.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
write.table(Y, file = 'data/Y_reg.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
write.table(causal_snps, file = 'data/causal_reg.txt', col.names = FALSE, row.names = FALSE,
quote = FALSE, sep = " ")
#RUN SUSIE
library(susieR)
fit = susie(X, Y)
fit$pip
fit$sets$cs