Artificial Intelligence in Drug Discovery and Design: A Comprehensive Review of Current Applications, Challenges, and Future Directions

Abstract

Objectives: This study seeks to critically examine how artificial intelligence (AI) is used in drug discovery and drug design and in particular it intends to examine how AI is used in the identification of a target, optimization of leads, de novo design and prediction of pharmacokinetic and toxicological properties.

Evidence acquisition: The major scientific databases, such as PubMed, Scopus, and Google Scholar were searched systematically and extensively to retrieve evidence. The peer-reviewed articles devoted to the usability of machine learning, deep learning, and generative artificial intelligence in pharmaceutical research were thoroughly reviewed and synthesized to offer a current overview of the latest advances, issues, and perspectives.

Results: AI technologies exert a tremendous influence on the sphere of drug discovery making possible the biological and chemical investigation of a big data. The strategies enhance the precision of forecasting, reduce the time it takes to develop the medicine and reduce the cost. It has been primarily used in drug-target interactions prediction, virtual screening, de novo molecular design and customized medicine. Such models as deep learning architectures, graph neural networks and generative models are further advanced to enhance molecular design and optimization. Although these benefits exist, other issues, such as the quality of data, model interpretability, legal concerns and integrating AI predictions into the experimentation process, are present.

Conclusion: Artificial intelligence has been a paradigm shift in the pharmaceutical research that has introduced new methods to offer medicines development more efficiently and quicker. As it continuously evolves, it ought to lead to the discovery of improved and safer medicines. Data quality, explainable AI models, and data integration between computational prediction and experimental validation could be some topics of future research.