SRPS: Survival Reinforced Transfer Learning for Multicentric Proteomic Subtyping and Biomarker Discovery

Omics-based molecular subtyping in large-scale and multicentric cohort studies is a prerequisite for proteomics-driven precision medicine (PDPM). However, maintaining subtypes with robust molecular features and significant prognostic associations across different cohorts remains challenging due to biological heterogeneity and technical inconsistency. Herein, we propose a subtyping algorithm, named Survival Reinforced Patient Stratification (SRPS), to adapt known subtypes from a discovery cohort to another by simultaneously preserving the distinct prognosis and molecular characteristics of each subtype. SRPS was benchmarked on simulated and real-world datasets, demonstrating a 12% increase in classification accuracy and best prognostic discrimination. Moreover, based on the calculated subtype significance score, an "unpopular" protein, peptidyl-prolyl cis-trans isomerase C (PPIC), was identified as the top 1 remarkable protein for subtyping hepatocellular carcinoma (HCC) patients with the worst prognosis. Eventually, PPIC was experimentally validated as a pro-cancer protein in HCC, confirming our work as a demonstration of interpretable machine learning-guided biological discovery in PDPM research. SRPS is publicly available at https://github.com/PHOENIXcenter/SRPS and https://ngdc.cncb.ac.cn/biocode/tool/BT007770.