Genome-wide Polygenic Risk Scores Predict Risk of Glioma and Molecular Subtypes

• DOI: 10.1101/2024.01.10.24301112
• Publication Date: 2024-01-11T14:10:12Z
• Authors: Taishi Nakase; Geno Guerra; Quinn T. Ostrom; Tian Ge; Beatrice Melin; Margaret Wrensch; John K. Wiencke; Robert B. Jenkins; Jeanette E. Eckel...; ; Melissa L. Bondy; Stephen S. Francis; Linda Kachuri
• Keywords: Polygenic risk score

ABSTRACT Background Polygenic risk scores (PRS) aggregate the contribution of many variants to provide a personalized genetic susceptibility profile. Since sample sizes glioma genome-wide association studies (GWAS) remain modest, there is need find efficient ways capturing factors using available germline data. Methods We developed novel PRS (PRS-CS) that uses continuous shrinkage priors model joint effects over 1 million polymorphisms on disease and compared it an approach (PRS-CT) selects limited set independent reach significance (P<5×10 -8 ). models were trained GWAS results stratified by histological (10,346 cases, 14,687 controls) molecular subtype (2,632 2,445 controls), validated in two cohorts. Results PRS-CS was consistently more predictive than PRS-CT across subtypes with average increase explained variance (R 2 ) 21%. Improvements particularly pronounced for glioblastoma tumors, yielding larger effect (odds ratio (OR)=1.93, P=2.0×10 -54 vs. OR=1.83, P=9.4×10 -50 higher =2.82% R =2.56%). Individuals 95 th percentile distribution had 3-fold lifetime absolute IDH mutant (0.63%) wildtype (0.76%) relative individuals PRS. also showed high classification accuracy mutation status among cases (AUC=0.895). Conclusions Our may improve identification high-risk help distinguish between prognostic subtypes, increasing potential clinical utility genetics patient management. IMPORTANCE OF THE STUDY Inherited variation one only few known contribute gliomagenesis. leverage largest collection show correlated genome yields improved prediction risk. improves according status. Additionally, we refined estimates individual highest percentiles confer significant increases Taken together, our findings further evidence genotyping be used as biomarker assessment non-invasive management patients newly diagnosed brain tumors.