Interpretable and generalizable quantitative models of relationships between chemical structures and their biological activities would enable informed efforts to optimize them. In my lab, we have prepared collections of small molecules characterized by high level of structural complexity and we captured some of their broad biological activity profiles through fluorescence imaging of stained cells. The challenge is to wade through this high-dimensional data and extract the meaningful connections between structure and activity.Scientific and engineering community has recently introduced new algorithms that may offer a way to address this difficult task. These algorithms search for sparse combinations of chemical features to predict the observed activity. Framing the qSAR problem as a search for such sparse solutions brings the toolbox of convex optimization to this important problem.

Robust qSAR models obtained from the intersection of new chemical space and cell morphology image space

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.