Finemapping through the FOCUS method (Mancuso et al., 2019) is a Bayesian statistical method designed to isolate subsets of genes more likely to contain causal genes based on a prior expectation of the number of causal genes we expect. We applied this approach to all suggestively significant regions in the AN GWAS (pGWAS < 1 × 10−5) to derive 90% credible sets. Credible sets were removed if they contained the null model that is included by FOCUS to account for missing causal mechanisms like genes without a suitable GReX model. Across either mRNA, protein or mRNA alternative splicing weights, there were 116 genes that were prioritised in a credible set (online Supplementary Table S5). Of these, eight genes (Table 1) had moderate (PIP > 0.4) evidence of a causal effect on AN, while five genes demonstrated strong evidence (PIP > 0.8). The strongest evidence of a causal relationship was observed for Neurexophilin And PC-Esterase Domain Family Member 1 (NXPE1) (PIP = 1, testis). Surprisingly, NXPE1 was indicated for the testis (GTEx), whilst it was also highly expressed in the colon
