The platelet-activating factor receptor (PAFR) is a G protein-coupled receptor (GPCR) that is implicated in multiple inflammatory response pathways as well as bacterial infection in the airway and several types of cancer. Thus, discovery of an effective antagonist against PAFR would lead to numerous therapeutic applications. However, the structure-function relationship of PAFR is poorly understood. In this project, we used a combination of docking, machine learning, and molecular dynamics (MD) simulations to characterize the structure-function relationship of PAFR and identify trends in formation of stable receptor-ligand complexes. MD simulations used Gaussian accelerated MD, in order to access conformations associated with the transition from the inactive to active state of PAFR. Small molecule docking was used with machine learning to massively reduce the computational time needed to screen ligand-receptor complexes. In particular, docking results utilized the Bag of Bonds (BoB) data feature to predict binding affinities, utilizing 20% of total data for training and subsequent testing against the rest of the ligand library. The top 5% of results predicted from BoB were cross-validated by running docking simulations on PAFR. Predicted docking energies were competitive with results from brute force docking. In combination with insights obtained from our MD simulations, we are able to show that ring-containing ligands are a necessary prerequisite for optimization of synthetic antagonists. Our approach holds promise for widespread application to other GPCRs.

Utilizing multi-scale modeling and machine learning approaches to characterize the structure-function relationship of the platelet-activating factor receptor

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.