A Computational Framework for Multi-Dimensional Drug Repurposing

ABSTRACT. Drug development and discovery represent one of the most demanding scientic domains, where rigorous validation and precision are required at every stage of the end-to-end pipeline. Even minor errors can result in substantial nancial loss and delays spanning several years. Beyond scientic and technical complexity, strict regulatory frameworks and legal constraints further intensify the burden, causing design, production, and commercialization processes to extend over long time horizons, often culminating in failure.

Where numerous diseases still lack validated therapeutic solutions and the majority of drug development trials do not succeed, the demand for ecient, scalable, and reliable discovery methodologies continues to grow alongside the global population. Although rapid technological advancements have signicantly improved scientic methodologies, many existing computational solutions remain impractical due to their hadware requirements. In practice, such requirements limit accessibility, hinder reproducibility, and constrain real-world adoption, particularly in resource-limited research environments. Consequently, the development of applicable, accurate, and accessible computational frameworks capable of operating under realistic hardware constraints remains a critical and unresolved challenge in modern drug discovery pipelines.

To address these challenges, we propose a computational methodology together with an open-source implementation that provides a practical and accessible solution to current limitations in drug discovery. The proposed approach operates on standard local computing environments, delivering high computational speed and accuracy, enabling efficient large-scale candidate evaluation. We aim to substantially enhance drug development and discovery pipelines, improving success rates by reducing both computational barriers and methodological constraints.