Drug ranking using machine learning systematically predicts the efficacy of anti-cancer drugs

This paper introduces Drug Ranking Using ML (DRUML), a machine learning approach using omics data to predict the efficacy of anti-cancer drugs. DRUML uses internally normalized distance metrics of drug response as features, reducing noise and improving predictive robustness. Trained on proteomics and phosphoproteomics data from 48 cell lines and verified with data from 53 cellular models from 12 independent laboratories, DRUML demonstrates low prediction error. Furthermore, its predictions of cytarabine sensitivity in clinical leukemia samples correlate with patient survival. The results suggest DRUML accurately ranks anti-cancer drugs across various pathologies.

Henry Gerdes, Pedro Casado, Arran Dokal, Maruan Hijazi, Nosheen Akhtar, Ruth Osuntola, Vinothini Rajeeve, Jude Fitzgibbon, Jon Travers, David Britton, Shirin Khorsandi, Pedro R. Cutillas