Integration of Mendelian randomization and transcriptome wide association studies for causal gene identification in ovarian cancer genetic architecture.

BACKGROUND: Ovarian cancer remains one of the most lethal gynecological malignancies with limited understanding of its genetic architecture and causal mechanisms. While genome-wide association studies have identified numerous susceptibility loci, the causal relationships between gene expression and ovarian cancer risk remain poorly understood. METHODS: We conducted a comprehensive genetic analysis integrating two-sample Mendelian randomization (MR) and transcriptome-wide association studies (TWAS) using expression quantitative trait loci (eQTL) data from the GTEx project and ovarian cancer GWAS data from the Ovarian Cancer Association Consortium (OCAC). Summary-data-based Mendelian randomization (SMR) analysis was performed to assess causal relationships between gene expression and ovarian cancer risk. Additionally, we analyzed survival outcomes in 3595 ovarian cancer patients from the SEER database to examine the clinical implications of radiation timing and constructed a prognostic nomogram incorporating genetic and clinical factors. RESULTS: MR analysis identified three genes with significant causal associations with ovarian cancer risk. GSTZ1 variant rs11621440 demonstrated a protective effect (OR = 0.999, 95% CI 0.999-1.000, FDR = 0.001), while PNP variant rs8015430 (OR = 1.006, 95% CI 1.002-1.010, FDR = 0.002) and KIAA1715 variant rs864697 (OR = 0.999, 95% CI 0.999-1.000, FDR = 0.002) showed increased risk associations. TWAS analysis revealed multiple genes with transcriptome-wide significant associations, including SIGLEC family genes (SIGLEC14, SIGLEC12, SIGLEC5) associated with increased risk through decreased expression, and metabolic genes (OCTN2, MSR1) associated with risk through increased expression. Survival analysis demonstrated that radiation timing significantly impacts prognosis, with postoperative radiation showing superior outcomes compared to preoperative radiation (HR = 1.49, 95% CI 1.23-1.82, p = 5.8 × 10-5). CONCLUSIONS: This study provides robust genetic evidence for causal relationships between specific gene expression patterns and ovarian cancer risk.