A Structured Mass Spectral Library of Food Contaminants Digitized from Official Regulatory Standards

It serves as a digital reconstruction of legally binding analytical protocols, bridging the gap between static regulatory documents and AI-ready data science.Current Scope (Version 1.0): The current release focuses on major global economies, integrating 18 authoritative official standards issued by China (GB standards), the United States (USDA/FDA/AOAC), and the European Union (CEN/EURL).

The data schema is designed to be scalable, supporting the future incorporation of standards from other jurisdictions (e.g., Japan, Canada) to build a global compliance knowledge graph.Data Content & Structure: The dataset contains 7,572 detection records covering 1,212 chemical entities (veterinary drugs, pesticides, and mycotoxins). It features "Contextual Fidelity," where mass spectral peaks are relationally linked to their specific validated experimental conditions (matrix types, sample preparation protocols, and instrument parameters) via a standardized JSON schema.Quality Assurance: To ensure the dataset serves as a standardized ground truth for regulatory method development, a 100% census audit was conducted against the original source documents, achieving a numerical accuracy rate of 99.8%.