by multiple correlated SNPs, and may cover multiple closely located genes. Some of these genes may be relevant to the disease, while others are not, yet due to the correlated nature of closely located genetic variants, distinguishing relevant from non-relevant genes is often not possible based on association P-values alone. Pinpointing the most likely relevant, causal genes and variants requires integrating available information about regional linkage disequilibrium (LD) patterns and functional consequences of correlated SNPs, such as deleteriousness of variants, but also their effects on gene expression as well as their role in chromatin interaction sites. Ideally, functional inferences obtained from different repositories are integrated, and annotated SNP effects are interpreted in the broader context of genes and molecular pathways. For example, consider a genomic risk locus with one lead SNP associated with an increased risk for a disease, and several dozen other SNPs in LD with the lead SNP that also show a low association P-value, spanning multiple genes. If none of these tested SNPs and none of the other (not tested but known) SNPs in LD with the lead SNP are known to have a functional consequence (i.e., altering expression of a gene, affecting a binding site or
