Multimodal Mixture-of-Experts with Retrieval Augmentation for Protein Active Site Identification

TL;DR

This paper introduces MERA, a retrieval-augmented multimodal framework with reliability-aware fusion for protein active site identification, achieving state-of-the-art results by effectively integrating multiple data modalities and estimating their trustworthiness.

Contribution

MERA is the first retrieval-augmented, multimodal mixture-of-experts framework with a reliability-aware fusion strategy for protein active site prediction, addressing data sparsity and modality reliability issues.

Findings

Achieves 90% AUPRC on active site prediction

Significant improvements on peptide-binding site identification

Validated effectiveness of retrieval-augmented multi-expert modeling

Abstract

Accurate identification of protein active sites at the residue level is crucial for understanding protein function and advancing drug discovery. However, current methods face two critical challenges: vulnerability in single-instance prediction due to sparse training data, and inadequate modality reliability estimation that leads to performance degradation when unreliable modalities dominate fusion processes. To address these challenges, we introduce Multimodal Mixture-of-Experts with Retrieval Augmentation (MERA), the first retrieval-augmented framework for protein active site identification. MERA employs hierarchical multi-expert retrieval that dynamically aggregates contextual information from chain, sequence, and active-site perspectives through residue-level mixture-of-experts gating. To prevent modality degradation, we propose a reliability-aware fusion strategy based on Dempster-Shafer evidence theory that quantifies modality trustworthiness through belief mass functions and learnable discounting coefficients, enabling principled multimodal integration. Extensive experiments on ProTAD-Gen and TS125 datasets demonstrate that MERA achieves state-of-the-art performance, with 90% AUPRC on active site prediction and significant gains on peptide-binding site identification, validating the effectiveness of retrieval-augmented multi-expert modeling and reliability-guided fusion.