Deep Learning Piloted Drug Design
Palak Singhal, Himanshu Bansal, Shweta Kumari, Sahil Jain · Bioinformatics and Computational Biology · 2023
Deep learning (DL) is a powerful machine learning technique based on artificial neural networks (ANNs), wherein multiple layers of processing are applied to extract progressively higher-level features from data. In recent years, a successful amalgamation of DL and bioinformatics approaches has been witnessed, especially in gene annotation, disease prediction, drug designing, and drug efficacy studies. Specifically for drug designing, synthetic technologies such as deep convolutional neural networks (DCNNs), recurrent neural networks (RNNs) and graph convolutional networks (GCNs) assist in expeditious development of therapeutic drugs. In this chapter, we discuss the application of different DL approaches in various drug design stages, namely, target identification and validation, lead optimization, virtual screening, predictive toxicology, clinical trial design and quantitative structure-activity relationship (QSAR). Further, we brainstorm on the advantages conferred by the application of DL in the drug designing process as compared to traditional approaches. Lastly, we discuss the future scope of DL in prediction, prevention, and treatment of various human ailments.