Automated Standardization of Legacy Biomedical Metadata Using an Ontology-Constrained LLM Agent

2026-03-10

arXiv.org

AI Summary

Biomedical metadata often suffers from incompleteness or noncompliance with community standards, undermining its discoverability, interoperability, and reusability. To address this challenge, this work proposes a large language model (LLM)-based approach for metadata standardization that innovatively integrates ontology-aware constraints with real-time tool invocation. By dynamically querying authoritative terminology services (e.g., OntoPortal) and machine-actionable metadata templates, the LLM retrieves up-to-date specifications on demand rather than relying solely on static training knowledge. Evaluated on 839 legacy HuBMAP records, the method significantly outperforms a baseline LLM-only strategy, achieving higher prediction accuracy and compliance across both controlled and uncontrolled fields.

Abstract

Scientific metadata are often incomplete and noncompliant with community standards, limiting dataset findability, interoperability, and reuse. When reporting guidelines exist, they typically lack machine-actionable representations. Producing FAIR datasets requires encoding metadata standards as machine-actionable templates with rich field specifications and precise value constraints. Recent work has shown that LLMs guided by field names and ontology constraints can improve metadata standardization, but these approaches treat constraints as static text prompts, relying on the model's training knowledge alone. We present an LLM-based metadata standardization system that queries authoritative biomedical terminology services in real time to retrieve canonically correct vocabulary terms on demand. We evaluate this approach on 839 legacy metadata records from the Human BioMolecular Atlas Program (HuBMAP) using an expert-curated gold standard for exact-match assessment. Our evaluation shows that augmenting the LLM with real-time tool access consistently improves prediction accuracy over the LLM alone across both ontology-constrained and non-ontology-constrained fields, demonstrating a practical, scalable approach to automated standardization of biomedical metadata.

Problem

• metadata standardization
• FAIR data
• biomedical metadata
• machine-actionable standards
• ontology constraints

Innovation

• ontology-constrained LLM
• real-time tool use
• metadata standardization
• FAIR data
• biomedical metadata