A Standardized Benchmark for Multilabel Antimicrobial Peptide Classification

2025-11-06

AI Summary

Antimicrobial peptide (AMP) research has long been hindered by fragmented data, inconsistent annotations, and the absence of standardized benchmarks. To address these challenges, we propose ESCAPE—a novel multi-label hierarchical annotation framework covering four antimicrobial activities: antibacterial, antifungal, antiviral, and antiparasitic. ESCAPE integrates 83,000 multi-source peptide sequences into a unified, reproducible large-scale benchmark dataset. Methodologically, we introduce a biologically semantically consistent hierarchical label taxonomy and design a Transformer-based model that jointly encodes sequence and structural features to enable simultaneous prediction of multiple activities and fine-grained antimicrobial spectrum identification. In multi-label classification, ESCAPE achieves state-of-the-art average precision, outperforming the second-best method by 2.56%. This work establishes the first comprehensive, standardized, and functionally granular AMP benchmark, advancing both computational AMP discovery and interpretable multi-activity prediction.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Multiagent Systems: Multiagent Planning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Agentic search
• Security and Privacy: Data transparency and provenance

Abstract

Antimicrobial peptides have emerged as promising molecules to combat antimicrobial resistance. However, fragmented datasets, inconsistent annotations, and the lack of standardized benchmarks hinder computational approaches and slow down the discovery of new candidates. To address these challenges, we present the Expanded Standardized Collection for Antimicrobial Peptide Evaluation (ESCAPE), an experimental framework integrating over 80.000 peptides from 27 validated repositories. Our dataset separates antimicrobial peptides from negative sequences and incorporates their functional annotations into a biologically coherent multilabel hierarchy, capturing activities across antibacterial, antifungal, antiviral, and antiparasitic classes. Building on ESCAPE, we propose a transformer-based model that leverages sequence and structural information to predict multiple functional activities of peptides. Our method achieves up to a 2.56% relative average improvement in mean Average Precision over the second-best method adapted for this task, establishing a new state-of-the-art multilabel peptide classification. ESCAPE provides a comprehensive and reproducible evaluation framework to advance AI-driven antimicrobial peptide research.

Problem

• Address fragmented datasets and inconsistent peptide annotations
• Establish standardized benchmark for antimicrobial peptide classification
• Improve multilabel prediction of peptide functional activities

Innovation

• Integrates over 80,000 peptides from 27 repositories
• Uses transformer-based model with sequence and structural data
• Establishes standardized multilabel hierarchy for peptide classification