A Systematic Study of Joint Representation Learning on Protein Sequences and Structures

TL;DR

This paper systematically explores joint protein representation learning by integrating language models with structure encoders, achieving state-of-the-art results in protein function prediction.

Contribution

It introduces a comprehensive framework combining sequence and structure encoders with novel fusion strategies, advancing protein representation learning.

Findings

Significant performance improvements over existing methods

Effective fusion strategies for sequence and structure data

New state-of-the-art in protein function annotation

Abstract

Learning effective protein representations is critical in a variety of tasks in biology such as predicting protein functions. Recent sequence representation learning methods based on Protein Language Models (PLMs) excel in sequence-based tasks, but their direct adaptation to tasks involving protein structures remains a challenge. In contrast, structure-based methods leverage 3D structural information with graph neural networks and geometric pre-training methods show potential in function prediction tasks, but still suffers from the limited number of available structures. To bridge this gap, our study undertakes a comprehensive exploration of joint protein representation learning by integrating a state-of-the-art PLM (ESM-2) with distinct structure encoders (GVP, GearNet, CDConv). We introduce three representation fusion strategies and explore different pre-training techniques. Our method achieves significant improvements over existing sequence- and structure-based methods, setting new state-of-the-art for function annotation. This study underscores several important design choices for fusing protein sequence and structure information. Our implementation is available at https://github.com/DeepGraphLearning/ESM-GearNet.