PPI2Text: Captioning Protein-Protein Interactions with Coordinate-Aligned Pair-Map Decoding

TL;DR

PPI2Text is a multimodal language model that generates detailed, free-text descriptions of protein-protein interactions from amino acid sequences, enhancing interpretability and integration with biological literature.

Contribution

It introduces a novel coordinate-aligned pair map decoding method and releases a large dataset for PPI captioning, advancing multimodal biological language modeling.

Findings

PPI2Text outperforms baselines on linguistic metrics.

It achieves higher factuality scores compared to synthesized references.

The model effectively captures interaction details across residue pairs.

Abstract

Protein-protein interaction (PPI) modeling has been widely studied as a binary or multi-label classification task. While emerging multimodal large language models (LLMs) can now describe single proteins, they remain unable to generate free-form descriptions of interactions between protein pairs. Moving beyond controlled vocabulary annotations, we propose to model PPI using free-text description, enabling richer expressiveness, improved interpretability, and better integration with literature knowledge base. We present PPI2Text, a multimodal LLM for free-form PPI captioning from amino acid sequences, that encodes each protein using ESM3 encoder, constructs a pair map from the two representations to capture interactions across all residue pairs, and autoregressively generates descriptions using a Qwen3 language decoder. We further introduce PaCo-RoPE, a coordinate-aligned positional encoding that aligns each axis of the pair grid with the residue positions of the corresponding protein. In addition, we release PPI2Text-Dataset, a 351k-pair corpus of free-form PPI descriptions aggregated from ten curated biological databases and further synthesized with Gemini under evidence-tiered prompting. PPI2Text consistently outperforms strong baselines across multiple ablation settings and evaluation protocols. It not only achieves higher scores on linguistic metrics against synthesized references, but also excels on factuality metrics, where an LLM-based judge evaluates outputs against raw biological evidence.