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

2026-03-02

AI Summary

This work proposes MERA, a novel framework addressing two critical challenges in protein active site prediction: the fragility of single-instance predictions due to sparse training data and performance degradation caused by unreliable modalities dominating multimodal fusion. MERA introduces, for the first time, a retrieval-augmented mechanism that dynamically aggregates contextual information through a hierarchical multi-expert retrieval system operating at the chain, sequence, and active site levels. It further integrates residue-level mixture-of-experts gating with a modality reliability assessment grounded in Dempster–Shafer evidence theory to enable robust multimodal fusion. Evaluated on the ProTAD-Gen and TS125 datasets, MERA achieves 90% AUPRC and significantly outperforms existing methods in peptide binding site identification, demonstrating its effectiveness and state-of-the-art performance.

Technology Category

• Machine Learning: Active Learning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

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.

Problem

• protein active site identification
• multimodal fusion
• modality reliability
• sparse training data
• residue-level prediction

Innovation

• retrieval-augmented learning
• mixture-of-experts
• multimodal fusion
• Dempster-Shafer evidence theory
• protein active site prediction