CoPeP: Benchmarking Continual Pretraining for Protein Language Models

TL;DR

This paper introduces CoPeP, a benchmark for evaluating continual pretraining methods on protein language models using decade-spanning datasets, showing that temporal information improves model performance and that continual learning methods can outperform naive approaches.

Contribution

We propose the CoPeP benchmark for continual pretraining of protein language models, including datasets, metrics, and evaluation of various continual learning methods.

Findings

Temporal meta-information improves perplexity by up to 7%.

Several continual learning methods outperform naive pretraining.

The benchmark enables large-scale evaluation of continual learning in protein modeling.

Abstract

Protein language models (pLMs) have recently gained significant attention for their ability to uncover relationships between sequence, structure, and function from evolutionary statistics, thereby accelerating therapeutic drug discovery. These models learn from large protein databases that are continuously updated by the biology community and whose dynamic nature motivates the application of continual learning, not only to keep up with the ever-growing data, but also as an opportunity to take advantage of the temporal meta-information that is created during this process. As a result, we introduce the Continual Pretraining of Protein Language Models (CoPeP) benchmark, a novel benchmark for evaluating continual learning approaches on pLMs. Specifically, we curate a sequence of protein datasets derived from the UniProt Knowledgebase spanning a decade and define metrics to assess pLM performance across 31 protein understanding tasks. We evaluate several methods from the continual learning literature, including replay, unlearning, and plasticity-based methods, some of which have never been applied to models and data of this scale. Our findings reveal that incorporating temporal meta-information improves perplexity by up to 7% even when compared to training on data from all tasks jointly. Moreover, even at scale, several continual learning methods outperform naive continual pretraining. The CoPeP benchmark offers an exciting opportunity to study these methods at scale in an impactful real-world application.