Seminar Series: Knowledge-Based and Generative Modeling Approaches to Computer-Aided Drug Discovery

Join the School of Data Science and Society for a seminar featuring Dr. Alex Tropsha, K.H. Lee Distinguished Professor, Division of Chemical Biology and Medicinal Chemistry at the Eshelman School of Pharmacy. The title of Dr. Tropsha’s talk is “Knowledge-Based and Generative Modeling Approaches to Computer-Aided Drug Discovery.”

Recent rapid growth of chemical and biomolecular databases has transformed the field of drug discovery into a branch of Big Data science. Naturally, this growth has created the demand for novel machine learning (ML) and artificial intelligence (AI) algorithms to accelerate the design and discovery of novel compounds with desired bioactivities and properties. I will discuss several recent methodologies and computational tools that address this overarching challenge. I will introduce the concept of biomedical knowledge graphs (KG) and discuss the development of the ROBOKOP (Reasoning Over Biomedical Objects linked in Knowledge-Oriented Pathways) KG. I will discuss the application of ROBOKOP to elucidate mechanistic pathways underlying drug therapeutic effects and discover new indications for known drugs (‘drug repurposing’). I will then discuss our studies on the development of the generative modeling approaches for de novo design of new chemical entities with desired properties (‘drug discovery’) that employ deep learning and accelerate both ligand- and target protein structure-based methods. Throughout the talk, I will emphasize the importance of careful data curation, model validation and hype-free assessment of the power of AI methods in chemical biology. Finally, I will discuss several emerging projects in diverse areas of research that illustrate the trans-disciplinary appeal of methods described in this presentation.