8-Oct-2026
BacSafe: A comprehensive database for integrated assessment of bacterial biosafety
International Society for Health Data Science
• BacSafe is a comprehensive, open-access bacterial biosafety database covering 20,630 species across 3,902 genera, with 89 biosafety-related attributes per entry compiled from 711,344 genome assemblies, over 2.5 million PubMed publications, and 20,437 food and pharmaceutical recall records.
• It introduces a dual-dimensional risk framework that decomposes bacterial biosafety into clinical risk and environmental risk, reflecting the fact that these two dimensions are not inherently correlated.
• It integrates empirical, real-world risk evidence, bridging the gap between genome-based theoretical prediction and actual risk in practice.
• A fine-tuned AI assistant (RAG + DeepSeek-R1) enables natural-language querying of the database.
• Decision-tree-based tools operationalize the assessment of "objectionable microorganisms" in non-sterile products and cleanroom environments, implementing the algorithms of the Chinese Pharmacopoeia (2025 edition) and PDA Technical Report 67, alongside genome-upload tools for species identification and phenotypic prediction.
- Funder
- Space Medical Experiment Project of CMSP, “Leading Goose” R&D Program of Zhejiang