Effective discrimination of Yersinia pestis and Yersinia pseudotuberculosis by MALDI-TOF MS using multivariate analysis.

Separating Yersinia pseudotuberculosis and Yersinia pestis is an important issue in plague diagnosis but can be extremely difficult because of the high similarity between the two species. MALDI-TOF MS has grown as a diagnostic tool with great potential in bacterial identification. Its application in this field is largely enhanced by multivariate analysis, especially in extracting subtle spectral differences. In this study, we built a complete MALDI-TOF MS data pipeline and found a Y. pestis-specific biomarker at 3063 Da closely related to Y. pestis plasminogen activation factor. Based on this, we achieved almost perfect separation between Y. pseudotuberculosis and Y. pestis (AUC = 0.999) using a supervised linear discriminant analysis (LDA) model. This is significantly better than the conventionally applied unsupervised spectral similarity comparison methods, such as hierarchical clustering analysis (HCA), which gave a separation accuracy of 75.0%. This new computing method paves the way for automatic differentiation between the two highly similar bacterial species with high separation accuracy.