Abstract

The complexity and high costs of traditional drug development have driven increased interest in drug repurposing, a strategy that explores new therapeutic uses for existing drugs. This approach leverages established safety profiles and can significantly accelerate the drug development process. Recent advances in computational methods, particularly those that utilize multi-omics data, have further enhanced the potential for systemic drug repurposing. In this study, we present ReMAP (Repurposing through Multi-Omics Analysis and Prediction), a model designed for predicting drug responses, with a focus on applications in cancer treatment. By integrating somatic mutations, copy number aberrations, and gene expression data, ReMAP outperforms traditional single-omics and early integration approaches in predictive accuracy. Utilizing data from the PRISM database, we identify potential new indications for drugs, including Dacomitinib for Head & Neck Squamous cell cancer. This work demonstrates the potential of multi-omics integration and machine learning to revolutionize drug repurposing in oncology, bridging the gap between computational predictions and clinical applications.