Abstract A057: Integrated proteogenomics uncovers ancestry-specific and shared molecular drivers in localized prostate cancer

Abstract Background: Black men experience higher incidence and mortality rates from prostate cancer, yet the molecular drivers of outcome disparities are not clearly defined. We performed a multi-omics analysis of localized PCa in the equal access Military Health System to identify ancestry-associated biological differences and prognostic markers. Methods: We analyzed tumors from 112 patients (57 Black, 55 White) using whole genome sequencing, quantitative proteomics, and phosphoproteomics. To identify ancestry-associated differences, we leveraged gene set enrichments from over-representation analyses and then used Multi-Omics Gene-Set Analysis (MOGSA) to determine the contribution of multi-omics features to these enrichments. Integrative clustering using iClusterBayes was applied to define molecular subtypes based on shared variation across multi-omics data. Germline regulation of the proteome was assessed via eQTL analyses. Prognostic significance was evaluated with Cox proportional hazards models (adjusted for age, PSA, and percent African ancestry); and ancestry-specific biomarker panels were validated using Kaplan–Meier and AUC analyses. Results: Black patients displayed higher genomic variability, stronger androgen response, fatty-acid metabolism, and EMT. White patients were enriched for DNA repair gene deletions, MYC/E2F signaling, mTORC1 activity, and cell-cycle progression. Phosphoproteomics revealed ancestry-specific kinase dependencies, with CK2α and CHEK2 activity elevated in Black patients, and CDK1/2, PRKD1, and CAMK2D/B activity in White patients. eQTL analyses highlighted germline regulation of the proteome independent of CNAs. MOGSA revealed 56 pathways grouped into five ancestry-associated subtypes, contrasting White-enriched apoptosis and epigenetic pathways with Black-enriched structural and DNA repair responses. Multiomics integration with iClusterBayes defined three reproducible molecular subtypes, externally validated, providing a basis for risk stratification with clinical potential. Somatic and germline regulatory differences converged on steroid hormone signaling, metabolic reprogramming, PI3K/AKT/mTOR, and DNA damage response pathways, with ancestry-associated immune and stromal signals. Prognostic analyses identified ancestry-specific and shared alterations associated with PCa progression. Prognostic CNA and protein panels exceeded PSA and pathology models (AUC >0.8), with ancestry-specific and universal biomarkers validated across cohorts. Conclusions: This integrative proteogenomics study of localized PCa defines ancestry-associated pathways, identifies reproducible molecular subtypes, and highlights germline regulatory influences on the tumor proteome. Ancestry-aware CNA, protein, and phosphoproteomic biomarker panels improve risk prediction and highlight potential therapeutic vulnerabilities, including kinase and metabolic dependencies. These findings provide a framework for refining ancestry-informed risk stratification and testing hypotheses of precision treatments to reduce outcome disparities. Citation Format: Cara C. Schafer, Tamara S. Abulez, Xijun Zhang, Kun-Lin Ho, Jiji Jiang, Denise Young, Jesse Fox, Kelly A. Conrads, Brian L. Hood, Gauthaman Sukumar, Darryl Nousome, Praveen-Kumar Raj-Kumar, Mariano Russo, Ayesha A. Shafi, Xiaofeng A. Su, Albert Dobi, Amina Ali, Sally Elsamanoudi, Jennifer Cullen, William D. Figg, Gyorgy Petrovics, Clifton L. Dalgard, Matthew D. Wilkerson, Nicholas W. Bateman, Thomas P. Conrads, Isabell A. Sesterhenn, APOLLO Research Network, Leigh Ellis, Craig D. Shriver, Gregory T. Chesnut, Shyh-Han Tan. Integrated proteogenomics uncovers ancestry-specific and shared molecular drivers in localized prostate cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Innovations in Prostate Cancer Research and Treatment; 2026 Jan 20-22; Philadelphia PA. Philadelphia (PA): AACR; Cancer Res 2026;86(2_Suppl):Abstract nr A057.