Cross-ancestry genome-wide association studies of liver function biomarkers uncover pleiotropic variants, systemic disease links and therapeutic targets

Liver function-related quantitative biomarkers (LFQBs) are essential for assessing hepatic health, yet prior genome-wide association studies (GWAS) have largely studied them in isolation. We conducted cross-ancestry GWAS meta-analyses on seven LFQBs to further elucidate liver function’s genetic architecture, identify pleiotropic variants, and prioritize genes to pinpoint targets with therapeutic potential. We performed GWAS meta-analyses on seven LFQBs in ~ 456,000 individuals across East Asian, European, South Asian, and African ancestries, followed by a series of downstream analyses, including fine-mapping, phenome-wide association study (PheWAS), gene-linking, and Mendelian randomization (MR). We identified 5,507 lead signals (P < 5 × 10−9), including 210 novel ones. Fine-mapping revealed 2,012 putative causal variants, of which 38 concurrently exhibited causal signals across multiple LFQBs and showed widespread associations with liver- and metabolism-related traits in PheWAS. Additionally, polygenic risk score (PRS)-based PheWAS uncovered pan-systemic manifestations of hepatic homeostasis disruption across diverse phenotypic domains. We proposed a novel multidimensional gene-linking framework (mdS2G) to bridge the identified causal loci to 1,166 putative genes. Benchmarking analysis revealed that mdS2G exhibited superior hepatocyte enrichment compared to individual constituent strategies. Furthermore, MR analysis highlighted PEPD protein as exhibiting therapeutic potential for metabolic dysfunction-associated steatotic liver disease (MASLD) and cirrhosis (P = 2.24 × 10−5, 4.36 × 10−3, respectively). This study maps the genetic landscape of liver function with expanded ancestral diversity, uncovering key variants and genes tied to hepatic homeostasis disruption and laying the groundwork for targeted liver disease therapies. Additionally, our open-access gene-linking framework provides a resource for the rapid prioritization of putative genes to guide future experimental interrogation.