Objective To establish a specific identification database for Streptobacillus moniliformis using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) technology, address the mis-identification of non-target strains from the same genus as S. moniliformis by current mass spectrometry databases, improve the accuracy of clinical and laboratory identification of this bacterial species.

Methods A total of 17 strains of S. moniliformis and 1 strain of S. notomytis isolated from clinical blood samples from 2019 to 2024 were identified by 16S rRNA sequencing method. The strains were then identified by using fully automated microbial identification instrument (Vitek-2 Compact) and fully automated microbial mass spectrometry identification system (Vitek-MS), respectively. Ten strains of S. moniliformis were selected and targets were collected by using Vitek-MS scientific research database, a map library was established by using software SARAMIS 4.17. The other 7 strains of S. moniliformis and 1 strain of S. notomytis were used to verify the self-built library. The clustering analysis results were compared based on self-built library and the Neighbor-joining phylogenetic relationship were compared based on 16S rRNA sequences of these 18 strains in this research. Additionally, S. notomytis was identified by mass spectrometry systems from different companies.

Results The 17 strains of S. moniliformis and 1 strain of S. notomytis were identified by 16S rRNA sequencing. Neither the Vitek-2 Compact nor the Vitek-MS (prior to library establishment) provided identification results for these 18 strains. In this study, a successful library was established by using 10 strains of S. moniliformis. The verification showed that the Vitek-MS self-built library could accurately identified 7 strains of S. moniliformis with a confidence level of 99.90%. However, no identification result of S. notomytis was obtained by using the S. moniliformis self-built library. Other mass spectrometry systems identified S. notomytis as S. moniliformis wrongly, the results were reliable at the species or genus levels. The results of clustering analysis by 16S rRNA and self-built library were consistent.

Conclusion This study successfully established a S. moniliformis mass spectrometry reference database by using Vitek-MS system. Preliminary validation demonstrated that this self-built database can effectively distinguish S. moniliformis from closely related species, significantly improving identification accuracy. However, given the limitations in strain quantity and diversity, further validation with additional isolates is required to confirm and optimize the database's performance.