Survey Paper on Various Techniques of Drug Discovery by Graph Neural Networks
Rounak Saha, Trishita Maity, Akanksha Singh, Gracy Kumari, Arnab Chakraborty · 2024
This survey report identifies and evaluates various approaches that could contribute to success in drug development and tracking activities using Graph Neural Networks (GNN). This paper aims to determine the best drug discovery approach among several related papers to increase the efficiency and the quality of the product. GNNs were incorporated into all levels of the drug discovery process, from target molecule discovery to de novo drug design. GNNs have significantly advanced efficiency and reliability in these processes, though challenges like biased datasets persist, creating room for improvement of the existing GNN structures. It discusses the curation and categorization of existing literature that has helped us grasp the concept of GNNs in drug discovery processes. This paper presents a detailed analysis and results of various applications of GNNs in reducing health risks, refining drug formulations, and drug discovery. A collective comparison of theories from distinguished research papers is proposed, involving discussions and literature surveys on the implementation of the GNN approach to date.