Performance of infrared spectroscopy for Salmonella spp. typing

International audience

English. Fourier-Transform infrared (FT-IR) spectroscopy was described as an alternative technique for bacteria typing. The aims of this study were, first, to evaluate the performance of FT-IR compared to PFGE (Pulsed Field Gel Electrophoresis) and MLST (MultiLocus Sequence Typing) for Salmonella spp. typing. Several parameters were considered i) the discriminating power of the three methods (FT-IR, PFGE, MLST), ii) the concordance of FT-IR with the two other typing methods, iii) the predictivity of FT-IR for PFGE and ST. Secondly, we evaluated the performance to identify potential markers associated with phenotypic characteristics of the strains (virulence and colonization). Fifteen strains of Salmonella from the three major serotypes in pork were considered: 5 S. enterica ser. Derby, 5 S. enterica ser. Typhimurium and 5 S. enterica ser. Typhimurium-monophasic variant. These strains, previously typed by PFGE and MLST, were selected to have similar and different ST. They were also tested for their virulence and colonization on Caco-2 and Ipec J2 cells, respectively, to identify markers associated to these characters. All strains were typed (3 runs with 5 replicates) using IR Biotyper® system from Bruker Daltonics. The spectra were processed by the IR Biotyper software, in the wavelength range corresponding to the polysaccharide region (1300-800 cm-1). Spectral data were displayed as dendrogram with a cutoff value calculated automatically by default, using Euclidean distance and average linkage according to the manufacturer's recommendations. The IR clusters were compared to PFGE clusters, ST, virulence and colonization levels with Comparing Partition website. Six IR clusters were identified, giving a Simpson’s index of diversity of 0.732 for FT-IR as for MLST (0.741) but lower than that of PFGE (0.883). Adjusted Rand index was used to evaluate the concordance of FT-IR with PFGE and MLST. The concordance was low, under 0.5. The predictivity analysis was realized through the adjusted Wallace index. Two strains in the same IR cluster had, respectively, 30% and 36% of chance to be in the same PFGE cluster or having a same ST. The concordance of FT-IR with level of virulence and colonization was low, under 0.4. For the predictivity analysis, two strains in the same IR cluster had, respectively, 13.6% and 33.1% of chance to have the same level of virulence or colonization.In conclusion, FT-IR could be used only as another method to type bacteria such as PFGE or MLST and not to predict ST, PFGE clusters and phenotypic characters.