ABSTRACT Bacterial contamination poses substantial threats to healthcare and industrial production; yet integrated resources for a comprehensive assessment of bacterial biosafety risks are currently limited. To bridge this gap, we developed BacSafe ( http://bacsafe.dmicrobe.cn/ ), a knowledge database that profiles the biosafety risks of bacterial species based on their environmental and clinical risk traits. Environmental risk traits, such as biofilm formation and tolerance to disinfectants, and clinical risk traits, including the presence of virulence and antimicrobial resistance genes, are curated from authoritative literature and genome predictions. In addition to the theoretical characteristics, BacSafe also encompasses empirical indicators reflecting real‐world risks posed by each bacterial species, including food and pharmaceuticals recall events, and document frequency in PubMed literature. Covering 20,630 species across 3902 genera, BacSafe supports natural language querying via a fine‐tuned artificial intelligence (AI)‐driven assistant that significantly outperforms general‐purpose large language models in accuracy and contextual relevance. Additionally, BacSafe provides decision tree‐based tools for evaluation of objectionable microorganisms in non‐sterile products and cleanroom environments. Taken together, by consolidating dispersed aspects of biosafety knowledge, BacSafe serves as a vital resource for bacterial risk assessment in clinical, industrial, and research settings, establishing a foundation for future AI‐assisted, data‐driven biosafety management.
Feng et al. (Sat,) studied this question.