If a GWAS reports summary statistics for separate subsamples (e.g., 23andMe and UK Biobank), should we meta-analyze them before using them in Genomic SEM?
Replying to “If a GWAS reports summary statistics for separate …”:
I’d suggest keeping them separate and then using genomic SEM to combine them in your model, as two indicators of the same latent trait. But be careful it’s not just the 10,000 SNPs for 23andme, because you need about 1 million hapmap3 SNPs for genomic SEM)
Replying to “If a GWAS reports summary statistics for separate …”:
Thank you!