Statistical fine-mapping aims to identify the causal variant(s) that are driving a GWAS association signal, enabling functional experiments to validate biological function. A straightforward approach to fine-mapping is to prioritize variants based on the strength of the marginal association statistics (i.e. ranking p-values)43. This is an effective strategy in the case of a single causal variant, but can be suboptimal when multiple causal variants are present, as the SNP with the top p-value at the locus may be tagging multiple causal variants. An alternative is to compute the posterior probabilities of causality for every SNP in the region, based on the likelihoods of the observed z-scores conditional on each possible set of causal variant(s)44. These posterior probabilities can be used to construct a credible set of SNPs, defined as the smallest set of SNPs that contains the true causal variant(s) with a given probability (typically 90% or 99%). Initial studies approximated the posterior probabilities of causality under a single causal variant assumption. Under this assumption, posterior probabilities of causality can be estimated from z-scores without the need for LD information45;
