A Balanced Hierarchical Multi-Label Classification Method for Protein Function Prediction

Le résumé fourni par la source

Protein function prediction is essential for enhancing our understanding of biological processes and advancing biology and medicine. Recent computational models have shown impressive accuracy in predicting protein sequences across hundreds or thousands of functional classes. However, these functions are often organized into a large, unbalanced hierarchical structure, like that defined by Gene Ontology, which can lead to prior errors and the neglect of rare functional classes. Many existing models also rely on intermediate protein structures for predictions, making them time-consuming and prone to inaccuracies for unknown protein. In this study, we introduce the MTP (Metric-learning then Pruning) model, which uses metric-learning and focuses on bottom-level annotations. We then implement a pruning step to exclude misclassified labels during metric-learning to enhance prediction accuracy and reliability. This article not only present the novel method, but also propose a thought of function prediction suitable on the structure like GO terms. Our validation shows that MTP significantly outperforms many contemporary models, particularly in predicting rare functional classes.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.