tooluniverse-gwas-finemapping

Query credible sets for the lead SNP or gene using OpenTargets_get_variant_credible_sets or gwas_search_snps. Variants with posterior probability above 0.5 are strong causal candidates; use functional annotations and eQTL evidence to break ties among variants in high LD.

The lead SNP has the most significant p-value but is often just the best-tagged variant on the genotyping array. The causal variant may be in perfect LD with it or be a non-coding regulatory variant, so always inspect the full credible set rather than assuming the lead SNP is causal.

The locus-to-gene (L2G) score from Open Targets predicts which gene a variant affects by integrating distance, eQTL evidence, chromatin interactions, and functional annotations. Scores above 0.7 indicate high confidence, 0.5-0.7 moderate, and below 0.3 low confidence.

Fine-mapping requires full GWAS summary statistics, an LD reference panel, sufficient signal strength (p < 5e-8), and computational resources. Studies lacking summary statistics or fine-mapping in Open Targets will not return credible sets.

Fine-mapping accuracy depends on LD structure matching the study population, and broad LD blocks limit resolution. Bayesian methods also depend on priors about causal variant counts, and results provide candidates requiring experimental validation, not proof.