Genome-wide meta-analysis across East Asian and European populations provides insights into the molecular basis of lung cancer

The genetic architecture of lung cancer (LC) varies substantially across populations. However, the limited ancestral diversity of existing discovery cohorts hampers comprehensive characterization of ancestry-specific risk mechanisms and limits the transferability of polygenic risk scores (PRSs) across populations. In addition, the interplay between germline variants and somatic mutations in LC remains poorly characterized. We performed a cross-ancestry genome-wide association meta-analysis (GWMA) studies comprising 53,650 LC cases and 486,210 controls of East Asian (EAS) and European (EUR) ancestry. PRSs of cross-ancestry were developed and validated in the UK Biobank (UKB) and China Kadoorie Biobank (CKB). Fine-mapping and expression quantitative trait locus (eQTL) colocalization were conducted to prioritize candidate susceptibility genes. Gene-set based PRSs were calculated based on pathway enrichment analyses, telomere and smoking-related GWASs and supporting literature, and their association with somatic mutations and prognostic metrics were studied. Functional validation of novel locus was performed to elucidate the mechanisms of LC. Cross-ancestry GWMA identified 36 independent variants across 30 loci, including one novel locus, and revealed remarkable differences in genetic correlations with smoking behaviors between East Asians and Europeans. PRS-CSx derived from the meta-analysis outperformed previous models in both the two biobanks (UKB: HR = 1.27, 95%CI: 1.23–1.31; CKB: HR = 1.26, 95%CI: 1.19–1.34). Susceptibility genes were grouped into four pathways: telomere maintenance, smoking behavior, epithelial cell biology, and DNA damage response. Corresponding gene-set-based PRSs were significantly associated with tumor mutation burden, mutational signatures, key driver mutations, and patient survival. Finally, functional assays further demonstrated that two variants at the novel locus at 16q23.3 regulate MPHOSPH6 expression and protein function, influencing telomere length and LC risk. Our study systematically elucidates the shared and distinct genetic architectures of LC between EAS and EUR populations, introduces PRS-CSx for effective LC risk stratification underscoring its clinical implications across diverse ancestries. Additionally, we uncover a novel mechanism by which functional genetic variants located in MPHOSPH6 contribute to LC susceptibility, and provide comprehensive insights into the interplay between germline and somatic mutations in LC.