Introduction: Colorectal cancer (CRC) ranks third as a leading cause of cancer-related death globally. The urgent development of accurate diagnostic and therapeutic strategies is extensively needed. The drug discovery process in oncology remains a critical bottleneck in modern medicine due to its high cost, long timelines, and frequent clinical trial failures. In contrast, Drug Repurposing, a subset of Drug Discovery, identifies new therapeutic uses for existing drugs and has emerged as a powerful strategy to accelerate treatment development. This study aims to establish a computational framework integrating transcriptomic profiling, network pharmacology, and machine learning to identify prognostic biomarkers and prioritize repurposable drugs for colorectal cancer (CRC).

Methods: Publicly available high-throughput RNA-seq data were analyzed to identify differentially expressed genes. A protein–protein interaction (PPI) network was constructed to isolate hub genes, of which three genes were significantly overexpressed in tumors and strongly correlated with poor patient survival. This process proved them to be key prognostic biomarkers. A drug–gene interaction network was built using curated databases. An unsupervised machine learning pipeline combining principal component analysis (PCA) and K-means clustering was developed to integrate gene expression data, survival scores, and interaction profiles for drug ranking.

Result: Among 34 candidate drugs, Palbociclib, Vorinostat, and Methotrexate were identified as high-potential multi-target drugs linked to all three identified potential biomarkers. These findings highlight multi-target repurposing opportunities in CRC.

Conclusion: This interdisciplinary approach demonstrates how omics data with AI-driven analytics can accelerate the discovery of personalized, multi-target therapies. The proposed framework offers a scalable, data-driven approach to rapidly identify drug candidates in colorectal cancer and other complex diseases.