Dynamic Molecular Graph-based Implementation for Biophysical Properties Prediction

TL;DR

This paper introduces a transformer-based GNN model that leverages dynamic 3D molecular data to improve predictions of biophysical properties and protein-ligand interactions, outperforming existing static models.

Contribution

It presents a novel dynamic GNN transformer architecture that incorporates time series data and physics-based simulations for enhanced molecular property prediction.

Findings

Outperformed existing models with an RMSE of 1.2958.

Utilized dynamic 3D data and geometric encodings.

Demonstrated the effectiveness of physics-based pre-training.

Abstract

Neural Networks (GNNs) have revolutionized the molecular discovery to understand patterns and identify unknown features that can aid in predicting biophysical properties and protein-ligand interactions. However, current models typically rely on 2-dimensional molecular representations as input, and while utilization of 2\3- dimensional structural data has gained deserved traction in recent years as many of these models are still limited to static graph representations. We propose a novel approach based on the transformer model utilizing GNNs for characterizing dynamic features of protein-ligand interactions. Our message passing transformer pre-trains on a set of molecular dynamic data based off of physics-based simulations to learn coordinate construction and make binding probability and affinity predictions as a downstream task. Through extensive testing we compare our results with the existing models, our MDA-PLI model was able to outperform the molecular interaction prediction models with an RMSE of 1.2958. The geometric encodings enabled by our transformer architecture and the addition of time series data add a new dimensionality to this form of research.