STAR-GO: Improving Protein Function Prediction by Learning to Hierarchically Integrate Ontology-Informed Semantic Embeddings

2025-12-04

AI Summary

Existing protein function prediction models typically exploit either the textual definitions or the hierarchical structure of Gene Ontology (GO) terms in isolation, limiting generalization to unseen GO terms and resulting in rapid model obsolescence. To address this, we propose the first multimodal framework that hierarchically fuses GO semantics—i.e., textual definitions—with its graph-structured hierarchy. Our approach employs a Transformer-based architecture to jointly encode protein sequences, GO term descriptions, and the GO directed acyclic graph, enabling semantic-structural alignment and cross-level information propagation. The method supports zero-shot GO function prediction and achieves state-of-the-art performance on standard benchmarks, particularly excelling in predicting functions for previously unseen GO terms. Code is publicly available.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Accurate prediction of protein function is essential for elucidating molecular mechanisms and advancing biological and therapeutic discovery. Yet experimental annotation lags far behind the rapid growth of protein sequence data. Computational approaches address this gap by associating proteins with Gene Ontology (GO) terms, which encode functional knowledge through hierarchical relations and textual definitions. However, existing models often emphasize one modality over the other, limiting their ability to generalize, particularly to unseen or newly introduced GO terms that frequently arise as the ontology evolves, and making the previously trained models outdated. We present STAR-GO, a Transformer-based framework that jointly models the semantic and structural characteristics of GO terms to enhance zero-shot protein function prediction. STAR-GO integrates textual definitions with ontology graph structure to learn unified GO representations, which are processed in hierarchical order to propagate information from general to specific terms. These representations are then aligned with protein sequence embeddings to capture sequence-function relationships. STAR-GO achieves state-of-the-art performance and superior zero-shot generalization, demonstrating the utility of integrating semantics and structure for robust and adaptable protein function prediction. Code is available at https://github.com/boun-tabi-lifelu/stargo.

Problem

• Improves protein function prediction by integrating ontology semantics and structure
• Addresses generalization to unseen Gene Ontology terms in evolving ontologies
• Enhances zero-shot learning by aligning protein sequences with hierarchical GO representations

Innovation

• Transformer-based framework for protein function prediction
• Integrates textual definitions with ontology graph structure
• Aligns GO representations with protein sequence embeddings