Application of a MALDI-TOF analysis platform (ClinProTools) for rapid and preliminary report of MRSA sequence types in Taiwan

• Main Authors: Hsin-Yao Wang, Frank Lien, Tsui-Ping Liu, Chun-Hsien Chen, Chao-Jung Chen, Jang-Jih Lu
• Format: Article
• Language: English
• Published: PeerJ Inc. 2018-11-01
• Series: PeerJ
• Subjects: MALDI-TOF; Lineage typing; MLST; ClinProTools; ML; MRSA
• Online Access: https://peerj.com/articles/5784.pdf

Background The accurate and rapid preliminarily identification of the types of methicillin-resistant Staphylococcus aureus (MRSA) is crucial for infection control. Currently, however, expensive, time-consuming, and labor-intensive methods are used for MRSA typing. By contrast, matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a potential tool for preliminary lineage typing. The approach has not been standardized, and its performance has not been analyzed in some regions with geographic barriers (e.g., Taiwan Island). Methods The mass spectra of 306 MRSA isolates were obtained from multiple reference hospitals in Taiwan. The multilocus sequence types (MLST) of the isolates were determined. The spectra were analyzed for the selection of characteristic peaks by using the ClinProTools software. Furthermore, various machine learning (ML) algorithms were used to generate binary and multiclass models for classifying the major MLST types (ST5, ST59, and ST239) of MRSA. Results A total of 10 peaks with the highest discriminatory power (m/z range: 2,082–6,594) were identified and evaluated. All the single peaks revealed significant discriminatory power during MLST typing. Moreover, the binary and multiclass ML models achieved sufficient accuracy (82.80–94.40% for binary models and >81.00% for multiclass models) in classifying the major MLST types. Conclusions A combination of MALDI-TOF MS analysis and ML models is a potentially accurate, objective, and efficient tool for infection control and outbreak investigation.