This package is for performing CD methods in the paper “Inferring causal direction between two traits in the presence of horizontal pleiotropy with GWAS summary data”.
Install development version from GitHub with:
# install.packages("devtools")
devtools::install_github("xue-hr/MRCD")We use LDL and CAD as an example.
library(MRCD)
## basic example codeThe list pruned contains two objects: loci_bed is the reference panel of 22 SNPs for 489 individuals, sig_part is the summary statistics for LDL and CAD:
pruned$loci_bed[1:6,1:6]
#> rs11206510 rs11591147 rs12740374 rs4245791 rs2351524 rs2282143
#> HG00096:HG00096 0 0 2 0 0 0
#> HG00097:HG00097 1 0 0 1 1 0
#> HG00099:HG00099 1 0 0 1 0 0
#> HG00101:HG00101 0 0 0 1 1 0
#> HG00102:HG00102 0 0 0 1 0 0
#> HG00103:HG00103 2 0 0 0 0 0
head(pruned$sig_part)
#> chr pos rsid A1 A2 beta_LDL se_LDL N_LDL beta_CAD se_CAD
#> 1 1 55496039 rs11206510 C T -0.0695200 0.003555 294565 -0.06272 0.01138
#> 2 1 55505647 rs11591147 T G -0.4752703 0.011494 265213 -0.22090 0.03500
#> 3 1 109817590 rs12740374 T G -0.1618858 0.003197 294565 -0.10990 0.01022
#> 4 2 44074431 rs4245791 C T 0.0721910 0.003093 270962 0.05444 0.00887
#> 5 2 203880992 rs2351524 T C -0.0239770 0.004204 295826 0.10227 0.01228
#> 6 6 160557643 rs2282143 T C 0.0582204 0.008825 247909 0.26186 0.03325
#> N_CAD loci
#> 1 336860 34
#> 2 268736 34
#> 3 268733 67
#> 4 268741 160
#> 5 268745 253
#> 6 243575 734Apply 3 CD methods to 22 SNPs from 12 loci:
set.seed(1)
results = CD_3_methods(pruned)Results of CD-Ratio:
results$CD_Ratio_result
#> $T1toT2
#> K se(K)
#> 0.145074866 0.006803127
#>
#> $T2toT1
#> K se(K)
#> 0.94092638 0.05439297
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 343.244
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 498.7441Results of CD-Egger:
results$CD_Egger_result
#> $T1toT2
#> b0 K se(b0) se(K)
#> 0.006055084 0.166677177 0.001990455 0.031729184
#>
#> $T2toT1
#> b0 K se(b0) se(K)
#> -0.02529651 3.29972149 0.00949699 0.63893717
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 22.49414
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 22.00414Results of CD-GLS:
results$CD_GLS_result
#> $T1toT2
#> b0 K se(b0) se(K)
#> 0.005959528 0.166556725 0.001965053 0.031333548
#>
#> $T2toT1
#> b0 K se(b0) se(K)
#> -0.02374490 3.11837442 0.00853066 0.58016321
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 22.72853
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 23.96457Now apply 3 CD methods to 12 indpendent SNPs, each of them from a locus. We do not need reference panel for independent SNPs.
set.seed(1)
sig_part = pruned$sig_part[c(2,3,4,5,6,9,11,13,14,15,18,21),]
independent_results = CD_3_methods_Independent(sig_part)Results of CD-Ratio:
independent_results$CD_Ratio_result
#> $T1toT2
#> K se(K)
#> 0.19413572 0.01050562
#>
#> $T2toT1
#> K se(K)
#> 0.72765268 0.06337732
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 184.0483
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 393.7104Results of CD-Egger:
independent_results$CD_Egger_result
#> $T1toT2
#> b0 K se(b0) se(K)
#> 0.008331157 0.220582429 0.002173359 0.040930430
#>
#> $T2toT1
#> b0 K se(b0) se(K)
#> -0.02831589 3.21230872 0.00918193 0.59210624
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 11.93857
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 11.99474Results of CD-GLS:
independent_results$CD_GLS_result
#> $T1toT2
#> b0 K se(b0) se(K)
#> 0.008162980 0.229319510 0.002289542 0.041957061
#>
#> $T2toT1
#> b0 K se(b0) se(K)
#> -0.028377515 3.005738170 0.008756814 0.521305169
#>
#> $Q_T1toT2
#> [,1]
#> [1,] 12.12634
#>
#> $Q_T2toT1
#> [,1]
#> [1,] 12.84648