Aim: Creation of a classification model of Leptospira spp. serovar model using ClinProTools 3.0 software and evaluation of use of MALDI-TOF MS as a method of quality control of reference strains of leptospira.
Materials and methods: 10 reference strains of Leptospira spp. were used in the study according to microscopic agglutination reaction from the collection of Pasteur RIEM. All the strains were cultivated for 10 days in Terskikh medium at 28 degrees C. Cell extracts were obtained by ethanol/formic acid method. α-cyano-4-hydroxycinnamic acid solution was used as a matrix. Mass-spectra were obtained in Microflex mass-spectrometer (Bruker Daltonics, Germany). External validation of the test-model was carried out using novel spectra of every reference strain during their repeated reseeding.
Results: Values of cross-validation and confirmatory ability of the optimal model, built on a genetic algorithm, was 99.14 and 100%, respectively. This model contained 11 biomarker peaks (m/z 2959, 3447, 3548, 3764, 3895, 5221, 5917, 6173, 6701, 7013, 8364) for serovar classification. Results of the external validation have shown a 100% correct classification in serovar classesin Sejroe, Ballum, Tarassovi; Copenhageni, Mozdoc, Grippotyphosa and Patoc, that indicates a high prognostic ability of the model in these classes. However, data from verification matrix have shown, that 50%.of the spectra from Canicola and Pomona serovars were classified as Patoc class, that could be associated with cross serological activity of Patoc serovar L. biflexa with pathogenic leptospirae.
Conclusion: MALDI-TOF mass-spectrometry method combined with building and using the classification model could be a useful instrument for intra-laboratory control of leptospira reseeding.