A systems-level machine learning approach uncovers therapeutic targets in clear cell renal cell carcinoma

Abstract Clear cell renal cell carcinoma (ccRCC) is an aggressive malignancy with limited treatment options and high rates of resistance to first-line kinase inhibitors. Current therapies largely target the tumor microenvironment, leaving intrinsic tumor vulnerabilities underexplored. Here, we introduce a systems-based machine learning pipeline that integrates single-cell RNA sequencing, protein interaction networks, and drug proximity analysis to identify therapeutic targets in ccRCC. Candidate genes were refined using CRISPR screening data and functional relevance and validated across independent transcriptomic datasets. The pipeline recovered several established treatment pathways and uncovered previously underexplored therapeutic mechanisms, including ABL1, CDK4/6, and JAK inhibition. We identified FDA-approved compounds acting through these pathways, three of which, Ribociclib, Ponatinib, and Dasatinib, showed superior efficacy to current therapies across renal cancer cell lines in preclinical screens. By acting through mechanisms distinct from current therapies, they represent promising candidates for combination strategies aimed at overcoming resistance and improving clinical outcomes in ccRCC.