forked from jmcole003/SDS_humans
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Perform_regression_SMA.R
162 lines (121 loc) · 5.31 KB
/
Perform_regression_SMA.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
# Perform SMA regression with weightings on 27 traits
## 5/2/24
# libraries
library(data.table)
library(dplyr)
#directory
dr <- "/scratch/ukb/data/"
#weighted RMA regression function
sma_weighted <- function(x, y, sx, sy) {
#weights
wx <- sx^2
wy <- sy^2
v <- 1/(wx + wy)
w <- v/sum(v)
#weighted means
x_bar <- sum(w * x)
y_bar <- sum(w * y)
#weighted covariance and variances
cov_xy <- sum(w * (x - x_bar) * (y - y_bar))
var_x <- sum(w * (x - x_bar)^2)
var_y <- sum(w * (y - y_bar)^2)
#slope and intercept
b1 <- sign(cov_xy)*(var_y / var_x)^0.5
b0 <- y_bar - b1 * x_bar
#SE
n <- length(x)
r <- cov_xy/sqrt(var_x*var_y)
se<- abs(b1) * sqrt((1-r^2)/n)
return(c(b0, b1,se))
}
#inputs
trait <- commandArgs(trailingOnly = TRUE)[1]
pexp <- commandArgs(trailingOnly = TRUE)[2]
pval_thres <- commandArgs(trailingOnly = TRUE)[3]
#trait file
filename_l <- paste0(dr, "traits/",trait,".gwas_ML_lowlfsr.txt")
GWAS_data_lfsr <- read.table(filename_l, header = TRUE)
#P-value threshold for GWAS snps
pval <- as.numeric(1)
#pval <- as.numeric(pval_thres)
GWAS_data_filtered_lfsr <- GWAS_data_lfsr[GWAS_data_lfsr$Male_lfsr <= pval | GWAS_data_lfsr$Female_lfsr <= pval,]
#Effect size sex differences
GWAS_data_filtered_lfsr$Effect_Diff <- (GWAS_data_filtered_lfsr$Female_pm - GWAS_data_filtered_lfsr$Male_pm)
#Effect size SEs
GWAS_data_filtered_lfsr$varEffect <- ((GWAS_data_filtered_lfsr$Female_psd)^2 + (GWAS_data_filtered_lfsr$Male_psd)^2)
GWAS_data_filtered_lfsr$sdEffect <- sqrt(GWAS_data_filtered_lfsr$varEffect)
#output summary GWAS file
write.table(GWAS_data_filtered_lfsr,
file = paste0(dr, "traits/",trait,".gwas_mash_summary.txt"),
sep = "\t",row.names = FALSE,append = FALSE)
################################################################################
#Bootstrap regression (viability)
slopes_V_l <- vector(length = 1000)
for(i in 1:1000) {
#randomly sample 1 snp per block
snp_sample_l <- GWAS_data_filtered_lfsr %>% group_by(Chrom,block) %>% slice_sample(n=1)
#fit model (SMA and OLS)
sma_via_l <- sma_weighted(snp_sample_l$Effect_Diff, snp_sample_l$s_v,
snp_sample_l$sdEffect, snp_sample_l$SE_v)
slopes_V_l[i] <- sma_via_l[2]
}
# viability out
Z_V_l <- mean(slopes_V_l)/sd(slopes_V_l)
boot_via_l <- data.frame("Trait" = trait,
"Mode" = "Viability",
"Data" = "Lowest LFSR",
"mean_slope" = mean(slopes_V_l),
"SD_slope" = sd(slopes_V_l),
"Z" = Z_V_l,
"sample_size_regression" = nrow(snp_sample_l),
"SNPs_after_pval_filtering" = nrow(GWAS_data_filtered_lfsr))
# get output file and write it
write.table(boot_via_l, file = paste0(dr, "traits/",trait,".regression.lowestlfsr.mash.",pexp,".v.result"),
sep = "\t", row.names = FALSE,append = FALSE)
###########################################################################
#Bootstrap regression (fecundity)
slopes_F_l <- vector(length = 1000)
for(i in 1:1000) {
#randomly sample 1 snp per block
snp_sample_l <- GWAS_data_filtered_lfsr %>% group_by(Chrom,block) %>% slice_sample(n=1)
#fit model (SMA and OLS)
sma_fec_l <- sma_weighted(snp_sample_l$Effect_Diff, snp_sample_l$s_f,
snp_sample_l$sdEffect, snp_sample_l$SE_f)
slopes_F_l[i] <- sma_fec_l[2]
}
# viability out
Z_F_l <- mean(slopes_F_l)/sd(slopes_F_l)
boot_fec_l <- data.frame("Trait" = trait,
"Mode" = "Fecundity",
"Data" = "Lowest LFSR",
"mean_slope" = mean(slopes_F_l),
"SD_slope" = sd(slopes_F_l),
"Z" = Z_F_l,
"sample_size_regression" = nrow(snp_sample_l),
"SNPs_after_pval_filtering" = nrow(GWAS_data_filtered_lfsr))
# get output file and write it
write.table(boot_fec_l, file = paste0(dr, "traits/",trait,".regression.lowestlfsr.mash.",pexp,".f.result"),
sep = "\t",row.names = FALSE,append = FALSE)
###########################################################################
#Bootstrap regression (total)
slopes_T_l <- vector(length = 1000)
for(i in 1:1000) {
#randomly sample 1 snp per block
snp_sample_l <- GWAS_data_filtered_lfsr %>% group_by(Chrom,block) %>% slice_sample(n=1)
#fit model (SMA and OLS)
sma_tot_l <- sma_weighted(snp_sample_l$Effect_Diff, snp_sample_l$s_t, snp_sample_l$sdEffect, snp_sample_l$SE_t)
slopes_T_l[i] <- sma_tot_l[2]
}
# viability out
Z_T_l <- mean(slopes_T_l)/sd(slopes_T_l)
boot_tot_l <- data.frame("Trait" = trait,
"Mode" = "Total",
"Data" = "Lowest LFSR",
"mean_slope" = mean(slopes_T_l),
"SD_slope" = sd(slopes_T_l),
"Z" = Z_T_l,
"sample_size_regression" = nrow(snp_sample_l),
"SNPs_after_pval_filtering" = nrow(GWAS_data_filtered_lfsr))
# get output file and write it
write.table(boot_tot_l, file = paste0(dr, "traits/",trait,".regression.lowestlfsr.mash.",pexp,".t.result"),
sep = "\t",row.names = FALSE,append = FALSE)