A survey on computational methods used for drug repositioning

Abstract

Drug repositioning offers a promising route to increase the efficacy and streamline the development of pharmaceutical treatments. With the high costs and extensive time frames associated with traditional drug development pathways, computational approaches have risen as pivotal alternatives. This paper surveys the latest network-based methodologies in computational drug repositioning, focusing on network analysis, machine learning, and deep learning methods. By examining the predictive accuracies of various methods across consistent datasets, we present a comparative analysis that reveals the strengths and potential of each category. Our findings highlight the evolution of computational strategies, emphasized by the growing complexity and predictive capabilities of recent models, especially those leveraging deep learning frameworks