PEER: A Comprehensive and Multi-Task Benchmark for Protein Sequence Understanding

TL;DR

PEER is a comprehensive multi-task benchmark for protein sequence understanding, evaluating various methods across diverse protein tasks and demonstrating the superiority of large-scale pre-trained models, with multi-task training enhancing performance.

Contribution

This paper introduces PEER, the first standardized multi-task benchmark for protein sequence understanding, enabling consistent evaluation of different deep learning approaches.

Findings

Pre-trained protein language models outperform traditional methods.

Multi-task learning improves performance across tasks.

Benchmark datasets and code are publicly available.

Abstract

We are now witnessing significant progress of deep learning methods in a variety of tasks (or datasets) of proteins. However, there is a lack of a standard benchmark to evaluate the performance of different methods, which hinders the progress of deep learning in this field. In this paper, we propose such a benchmark called PEER, a comprehensive and multi-task benchmark for Protein sEquence undERstanding. PEER provides a set of diverse protein understanding tasks including protein function prediction, protein localization prediction, protein structure prediction, protein-protein interaction prediction, and protein-ligand interaction prediction. We evaluate different types of sequence-based methods for each task including traditional feature engineering approaches, different sequence encoding methods as well as large-scale pre-trained protein language models. In addition, we also investigate the performance of these methods under the multi-task learning setting. Experimental results show that large-scale pre-trained protein language models achieve the best performance for most individual tasks, and jointly training multiple tasks further boosts the performance. The datasets and source codes of this benchmark are all available at https://github.com/DeepGraphLearning/PEER_Benchmark