Protein binding affinity prediction under multiple substitutions applying eGNNs on Residue and Atomic graphs combined with Language model information: eGRAL

TL;DR

This paper introduces eGRAL, a novel SE(3) equivariant graph neural network that predicts protein binding affinity changes due to multiple substitutions by integrating multi-scale features and large language model data, trained on a large simulated dataset.

Contribution

The paper presents eGRAL, a new architecture combining residue, atomic, and evolutionary features with language models for affinity prediction, trained on a large synthetic dataset.

Findings

eGRAL achieves accurate affinity change predictions.

The model benefits from large-scale simulated training data.

It outperforms existing methods on experimental datasets.

Abstract

Protein-protein interactions (PPIs) play a crucial role in numerous biological processes. Developing methods that predict binding affinity changes under substitution mutations is fundamental for modelling and re-engineering biological systems. Deep learning is increasingly recognized as a powerful tool capable of bridging the gap between in-silico predictions and in-vitro observations. With this contribution, we propose eGRAL, a novel SE(3) equivariant graph neural network (eGNN) architecture designed for predicting binding affinity changes from multiple amino acid substitutions in protein complexes. eGRAL leverages residue, atomic and evolutionary scales, thanks to features extracted from protein large language models. To address the limited availability of large-scale affinity assays with structural information, we generate a simulated dataset comprising approximately 500,000 data points. Our model is pre-trained on this dataset, then fine-tuned and tested on experimental data.