A research was undertaken into the application of Fourier transform infrared spectroscopy and artificial neural networks to identify bacteria of the Lactococcus genus at a level of species and subspecies. A methodology of FTIR spectra measurement and strategy of their analysis, previously developed by the author, were used to select FTIR spectra of bacteria strains analyzed and to expand the FTIR library. The research comprised 7 reference strains of milk fermentation bacteria of the Lactococcus genus derived from Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) and 88 strains isolated from food products. The bacteria strains of the Lactococcus genus were isolated from raw milk and from fermented milk products. The FTIR spectra of strains compiled in the library were used to develop artificial neural networks. The application value of selected artificial neural networks was verified based on the identification results of reference strains, as well as of the strains isolated from food products; their species affiliation was determined based on a PCR test using species-specific primers. The application of artificial neural networks to analyse FTIR spectra made it possible to correctly identify bacteria of the Lactococcus genus belonging to a particular species.
Lactococcus, FTIR spectroscopy, artificial neural networks