Retrieval-Augmented Multimodal Learning for Enzyme-Substrate Interaction Prediction Under Low-Homology Shift
2026-06-22 • Machine Learning
Machine Learning
AI summaryⓘ
The authors developed a method called RAMMESI to predict how enzymes and substrates interact, which is important for discovering new biocatalysts. Their method works well even when there is little similar data available and when test enzymes are quite different from the training examples. They do this by comparing the enzyme and substrate together, plus using related enzymes at test time to improve predictions. Their approach also handles imbalanced data better during training. Tests showed RAMMESI performs strongly, especially when enzymes have low similarity to known ones, and its retrieval component can help other models too.
enzyme substrate interactionbiocatalystsequence identitydistribution shiftcross-modal interactionweighted binary cross-entropyretrieval augmentationlow-homologycontextual aggregation
Authors
Chen Liu, Bingxin Zhou, Xinyuan Wang, Ming Li, Guisheng Fan, Liang Hong
Abstract
Enzyme substrate interaction (ESI) prediction is a fundamental computational task for biocatalyst discovery and reaction screening in large biochemical spaces. In practical settings, ESI prediction is challenged by sparse positive supervision and low-homology distribution shift, where test enzymes share limited sequence identity with those observed during training. To address these challenges, we propose RAMMESI, a retrieval-augmented multimodal framework for robust ESI prediction. RAMMESI learns explicit pairwise enzyme-substrate representations through directional cross-modal interaction modeling and adaptive fusion. To enhance robustness, RAMMESI retrieves neighboring enzymes at inference time, recombines them with the query substrate, and aggregates the resulting pairwise predictions as contextual evidence. To improve learning under sparse positive supervision, we further adopt an imbalance-aware weighted-BCE objective. Experiments on two ESI benchmarks under sequence-identity-aware splits demonstrate that RAMMESI achieves consistently strong performance, with particular advantages in more challenging low-identity regimes. In addition, the retrieval module improves multiple ESI backbones in a plug-and-play manner, suggesting that retrieval provides a general mechanism for improving robustness under homology shift.