Interleaved Tool-Call Reasoning for Protein Function Understanding

TL;DR

This paper introduces PFUA, a tool-augmented reasoning model for protein function prediction that leverages external biological tools, significantly outperforming traditional text-based reasoning models.

Contribution

The paper proposes PFUA, a novel framework integrating domain-specific tools into protein reasoning, addressing limitations of text-only models in biological knowledge application.

Findings

PFUA outperforms text-only models by 103% on average.

Tool integration improves reasoning accuracy and verifiability.

External biological priors enhance generalization in protein function prediction.

Abstract

Recent advances in large language models (LLMs) have highlighted the effectiveness of chain-of-thought reasoning in symbolic domains such as mathematics and programming. However, our study shows that directly transferring such text-based reasoning paradigms to protein function understanding is ineffective: reinforcement learning mainly amplifies superficial keyword patterns while failing to introduce new biological knowledge, resulting in limited generalization. We argue that protein function prediction is a knowledge-intensive scientific task that fundamentally relies on external biological priors and computational tools rather than purely internal reasoning. To address this gap, we propose PFUA, a tool-augmented protein reasoning agent that unifies problem decomposition, tool invocation, and grounded answer generation. Instead of relying on long unconstrained reasoning traces, PFUA integrates domain-specific tools to produce verifiable intermediate evidence. Experiments on four benchmarks demonstrate that PFUA consistently outperforms text-only reasoning models with an average performance improvement of 103%.