Home Department: undeclared
Mentor: Ehsan Adeli (Psychiatry & Behavioral Sciences - Public Mental
Health & Population Sciences)

“Autoencoder-Based Latent Space Exploration of Compound-Target Interactions Towards Drug Discovery Simulation"

Drug discovery can be expensive and time-consuming, with costs up to $4.46 billion in R&D and development reaching upwards 15 years. Compound-target simulations can reduce costs for the discovery period by narrowing down initial target and compound selections. Alexis’s project will utilize an autoencoder trained on bioactivity data on compound-target interactions to extract predictions about target properties (binding site flexibility, druggability, stability) from the latent space. Predicted properties can be used to simulate and visualize targets by integrating with existing molecular dynamic simulators. The goal is to optimize target identification and in silico compound discovery.