I was looking at Combined-PC for GWAS, and their method (if I understand) says that since Principal Components are independent, you can just add the test statistics for the PCs together to get a combined test statistic for each SNP.
I am in a situation where I have longitudinal data for a single phenotype I am reducing with fPCA to a set of PCs (and wish to include enough PCs to capture >95% variance explained). Does SAIGE's Tstat, support the same logic? Can I just run SAIGE on the PCs individually, then sum Tstats to get a combined Tstat for each SNP?
Hello,
I was looking at Combined-PC for GWAS, and their method (if I understand) says that since Principal Components are independent, you can just add the test statistics for the PCs together to get a combined test statistic for each SNP.
I am in a situation where I have longitudinal data for a single phenotype I am reducing with fPCA to a set of PCs (and wish to include enough PCs to capture >95% variance explained). Does SAIGE's Tstat, support the same logic? Can I just run SAIGE on the PCs individually, then sum Tstats to get a combined Tstat for each SNP?
Thanks, Rob