Morgan McMurray, Dr. Thomas Schulze, & Dr. Jennifer Brown
The application of machine learning in 21st century drug discovery and design is one of the most exciting areas of contemporary biopharmaceutical R&D, promising generation of lead molecule structures with less time, less money, and greater accuracy than traditional approaches. These promises are particularly attractive for drug development focused on the central nervous system, an area both critically in need of novel therapeutics and notoriously difficult to target, largely due to how challenging it is to design compounds that successfully pass through the blood-brain barrier (BBB) and/or the blood-cerebrospinal fluid (CSF) barrier. In an effort to help address this obstacle, my research applies Graph Neural Networks (GNNs) to generate novel chemical compounds that can enter the BBB, the blood-CSF barrier, or both. GNNs are deep learning algorithms designed to interpret data represented as mathematical graphs — a collection of nodes connected by edges — which are very intuitive structures to represent molecules in a computer.