Artificial Intelligence and Machine Learning for Exploring PROTAC in Underutilized Cells

PROTACs consist of a protein-targeting ligand chemically linked to an E3 ligase ligand. PROTACs facilitate the ubiquitination and degradation of target proteins by utilizing E3 ligases. The capacity of PROTACs to selectively induce protein breakdown, as opposed to inhibition, presents novel opportunities for tackling diseases linked to proteins that are difficult to address using conventional small-molecule inhibitors. Designing PROTAC molecules involves a careful selection of the recognition element, linker, and ubiquitin ligase binding element to ensure specificity, efficacy, and minimal off-target effects. Computational methods, structure-activity relationship studies, and iterative optimization are often employed in the design process. This opens newer avenues for utilizing the rapidly evolving Artificial Intelligence in PROTAC development. Recent exploration methods have mostly been centered around: (a) Using molecular docking to precisely model the protein-PROTAC-E3 ternary complex for PROTACs. This approach enables quick study of structure-activity relationships and enhances ligand affinity and selectivity. Virtual screening enables researchers to uncover potential ligands for the E3 ligase and target protein. (b) Generative methods explore chemical space to generate novel PROTAC structures. Deep learning models assist in predicting ligand-protein interactions and optimizing drug properties, and (c) The stability of the ternary complex generated by PROTACs, target proteins, and E3 ligases is assessed using AI-driven Molecular Dynamics (MD) simulations. In summary, AI and ML accelerate PROTAC drug discovery by enhancing ligand selection, predicting interactions, and optimizing drug properties. This chapter explores the transformative potential of integration of PROTAC technology and the rapidly advancing field of AI-ML in drug discovery.