Promising potential of machine learning-assisted MALDI-TOF MS as an effective detector for Streptococcus suis serotype 2 and virulence thereof

Abstract

ABSTRACT Streptococcus suis is a major swine pathogen that can be transmitted to humans by close contact with infected or carrier pigs. Among the 29 true serotypes, serotype 2 is the most significant due to its high prevalence in diseased pigs and its association with human clinical cases. The co-existence of strong and weak virulent strains of SS2 ( Streptococcus suis serotype 2) complicates the surveillance, coupled with the fact that traditional methods based on serologic tests or multiple PCR are laborious and time-consuming, posing challenges to the prevention and control of streptococcosis. Therefore, developing a new method to rapidly detect S. suis serotype 2 and virulent strains thereof is extremely essential. In this study, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) was performed among 137 S . suis strains covering all the true serotypes to explore its potential for identifying SS2 strains with strong virulence. By comparing mass spectrometry data with those of bacterial phenotypic and genomic properties, our study reveals, for the first time, the strong correlations between mass spectra patterns and the phenotypes (of serotype and virulence). In addition, machine learning-based classifiers have more than 90% accuracy in making correct judgments. Overall, the promising potential of MALDI-TOF MS for virulent SS2 detection was proven, and the corresponding methods disclosed in this study will pave the way for high speed-specific causative agent identification among more extensive bacterial species. IMPORTANCE To the best of our knowledge, this study reveals a strong correlation between mass spectra pattern and virulence phenotype among S. suis for the first time. In order to make the findings applicable and to excavate the intrinsic information in the spectra, the classifiers based on the machine learning algorithms were established, and RF (Random Forest)-based models have achieved an accuracy of over 90%. Overall, this study will pave the way for virulent SS2 ( Streptococcus suis serotype 2) rapid detection, and the important findings on the association between genotype and mass spectrum may provide a new idea for the genotype-dependent detection of specific pathogens.