GLProtein: Global-and-Local Structure Aware Protein Representation Learning

2025-05-17

arXiv.org

AI Summary

Existing protein representation methods struggle to jointly model global structural similarity and local amino acid-level details, limiting functional prediction accuracy. To address this, we propose the first end-to-end, cross-scale protein structure-aware learning framework that unifies atomic-level 3D distance modeling, substructure-level molecular encoding, and inter-protein structural similarity ranking. Our method introduces a joint optimization paradigm combining masked language modeling with triplet-based structural scoring, integrating contrastive learning and self-supervised pretraining. Extensive experiments demonstrate significant improvements over state-of-the-art methods on contact map prediction, protein–protein interaction prediction, and other downstream tasks. Ablation studies confirm that the synergistic modeling of global and local structural information substantially enhances structural awareness and generalization capability across diverse protein-related tasks.

Technology Category

• Machine Learning: Structured Learning
• Search and Optimization: Learning to Search
• Computer Vision: Representation Learning for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Proteins are central to biological systems, participating as building blocks across all forms of life. Despite advancements in understanding protein functions through protein sequence analysis, there remains potential for further exploration in integrating protein structural information. We argue that the structural information of proteins is not only limited to their 3D information but also encompasses information from amino acid molecules (local information) to protein-protein structure similarity (global information). To address this, we propose extbf{GLProtein}, the first framework in protein pre-training that incorporates both global structural similarity and local amino acid details to enhance prediction accuracy and functional insights. GLProtein innovatively combines protein-masked modelling with triplet structure similarity scoring, protein 3D distance encoding and substructure-based amino acid molecule encoding. Experimental results demonstrate that GLProtein outperforms previous methods in several bioinformatics tasks, including predicting protein-protein interaction, contact prediction, and so on.

Problem

• Integrating global and local protein structural information for representation learning
• Enhancing protein function prediction accuracy through structural awareness
• Addressing limitations in current protein sequence analysis methods

Innovation

• Combines masked modeling with triplet similarity scoring
• Integrates 3D distance encoding for structural representation
• Uses substructure-based amino acid molecule encoding