Rapid identification of ST17 group B streptococcus using MALDI-TOF MS / 中华检验医学杂志

ABSTRACT

Objective:To establish a classification model for rapid identification of hypervirulent subtype ST17 clones of Group B Streptococcus (GBS) using matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS).Methods:In a retrospective study, 235 strains of GBS strains were selected from multiple centers in China during 2015-2018. For model generation,45 strains of ST17 and 50 strains of non-ST17 (20 ST19, 15 ST12 and 15 ST10 strains) were enrolled as the modeling group. The remaining 90 main ST strains (40 ST17, 16 ST10, 17 ST12 and 17 ST19) were served as validation group. 50 GBS strains classified as other minor ST subtypes were regarded as taxonomic groups. MS spectra were collected by Bruker mass spectrometry, and then loaded for model generation and verification, and screening of differential peptide peaks by genetic algorithm (GA) and model verification on ClinProTools 3.0 software.Results:The recognition rate for ST17-GA model were 99.4% with cross validation value of 96.9%. Among the ten differential peptide peaks for the classification model, the weights of both two main peptide peaks m/z 2 956 and m/z 5 912 were greater than 1, while the weights of the all other eight peptide peaks were less than 0.5. Model validation showed only one of the ST17 was misjudged as non-ST17 strain, resulting in diagnostic accuracy of 98.9%, sensitivity of 97.5% and specificity of 100%, positive predictive value of 100% and negative predictive value of 98.0%, respectively. For other sporadic STs, 42.0% (21/50) of them were misdiagnosed as ST17 subtype.Conclusion:A MALDI-TOF MS classification model for hypervirulent subtype of ST17 GBS strains has been successfully established with good diagnostic efficacy.