The Locus Overlap Score Can Be Used for Prioritizing Trait-Associated Loci
After the most significantly enriched annotations are identified, the loci contributing the most to such enrichment are those with the lowest overlap score (Figure S8). A locus that obtains a low overlap score would typically have only a few variants linked to the index SNP and sparse density of the annotation. These are the loci where variants might be most effectively prioritized.
For example, in breast cancer, the locus with the best (lowest) overlap score (0.097) was rs889312 (Figure 6). That SNP is in LD with seven other variants, of which only rs1862626 overlaps a vHMEC H3K4me1 summit region. This SNP is upstream of MAP3K1 (MIM: 600982) and modifies a predicted binding site for estrogen receptor alpha (ER-α),50 consistent with the well-established role of estrogen-mediated signaling in breast cancer progression.51,52 For height, rs11677466 showed the best overlap score (0.026; Figure 6). This SNP itself overlaps an embryonic stem cell DHS that is a known binding site for HNF4α,53–55 a transcription factor that plays important roles in metabolic regulation and stem cell differentiation. Functional follow-up will be necessary to further validate these variants. The specific annotations indicate the type of regulatory element driving these associations and define the cell type in which they are active.