11.5 Expression QTLs

The variants we identify with this GWAS approach are called expression quantitative trait loci (eQTLs).

Notably, because the number of phenotypes we have is equal to the number of genes we’re testing, we would have to perform ~20,000 GWASs (the number of gemes) for every eQTL study in humans, and test ~5 million SNPs per GWAS.

Multiple testing in eQTL studies

When we’re performing this many repeated tests, you run the risk that any hits you find are just due to chance.

In a GWAS, we would use multiple testing correction to account for this. However, because of the overwhelming number of tests we need to perform to identify eQTLs, our significance threshold would be so low that we likely wouldn’t discover anything.

Reducing the number of tests

In order to reduce the number of tests, we typically focus on eQTL mapping just in in the region around the gene body, which dramatically decreases the number of SNPs we’re testing.

The rationale for this approach is that nearby regulatory regions (promoters, enhancers, etc.) are more likely to influence expression.

Fig. 4. eQTL studies typically restrict their QTL search space to variants within 1 Mb of the gene.