Objective To construct a database for molecular mass spectra of toxins and drugs in order to facilitate the management and retrieval of mass spectra for nerve agents,metabolites and other small molecules.Methods Requirement analysis and functional design were performed using software engineering methods.The Spec2Vec algorithm was used for vector representation of mass spectra,while SMILES molecular structures were vectorized using the extended connectivity fingerprint(ECFP).A data storage model integrating structured information and vector representations was established using the Milvus database.Similarity search of mass spectra and molecular structures was conducted via vector similarity comparison and the FlashEntropySearch algorithm.Results The constructed database of mass spectra encompassed over 400,000 entries from such sources as OCAD,NIST,MASSBANK,metabolic products,and natural products of TCM,which was capable of searching for similarities in mass spectra and molecular structures.On a standard server,similarity search of mass spectra took no more than 5 seconds,while that of molecular structures took no more than 1 second.Conclusion The system enables efficient management of complex mass spectra and provides rapid retrieval and comparison of mass spectra-related information through advanced vector indexing technology,offering robust data support and research tools for toxicology and pharmacology.

Androgenetic alopecia is a common clinical disease. The pathogenesis of this disease is not clear, while TCM has achieved sound therapeutic effects on this disease by syndromes differentiation. By studying theory, the pathogenesis, clinical treatment and laboratory research of TCM treating the androgenetic alopecia in the past 10 years, we tried to provide reference for the treatment of androgenetic alopecia in Chinese medicine.