Thanks for the fantastic lectures. I had previously asked this question during the lecture but need a reiteration including helpful resources as I am new to this kind of analysis “Is it correct practice to conduct eQTL analysis on a different ancestry when the SNPs selected were based on another ancestry”
If this kind of analysis is acceptable, what is the best way to go about it?
If the variants are ascertained in another population then you should expect to loose information because the frequency of these variants will be (expected) lower in your target population. Therefore, your analysis will be suboptimal but still “correct” in principle.
interesting insight, thank you so much for the fedback
@Ofure_U @l.yengo One extra point here I’ve been thinking about on this. eQTL reference panels are increasing constructed over larger more diverse cohorts. Because the actual physical effect of a particular SNPs allele on a gene’s expression will not be truly different across ancestry, an eQTL in reality is always pan ancestry (hence why studies like gTEX found differences in just a couple of genes). The difficulty then is just in detecting them, like Dr. Yengo said, if your on hand cohort is primarily one ancestry and your eQTL reference did not include at least enough participants with the ancestry unique alleles to sufficiently detect an eQTL then you won’t be able to say anything about that gene.
Dr. Yengo, what do you think of this? Am I missing some other issue with cross-ancestry eQTLs?
@JamesHart this is certainly one important information to keep in mind during the analysis. Thanks so much