Application of the Random Forest method to analyse epidemiological and phenotypic characteristics of Salmonella 4,[5],12:i:- and Salmonella Typhimurium strains.

Salmonella enterica 4,[5],12:i:- is a monophasic variant of S. Typhimurium. In the last decade, its prevalence rose sharply. Although S. 4,[5],12:i:- and S. Typhimurium are known to pose a considerable public health risk, there is no detailed information on the circulation of these serovars in Italy, particularly as far as veterinary isolates are concerned. For this reason, a data set of 877 strains isolated in the north-east of Italy from foodstuffs, animals and environment was analysed during 2005-2010. The Random Forests (RF) method was used to identify the most important epidemiological and phenotypic variables to show the difference between the two serovars. Both descriptive analysis and RF revealed that S. 4,[5],12:i:- is less heterogeneous than S. Typhimurium. RF highlighted that phage type was the most important variable to differentiate the two serovars. The most common phage types identified for S. 4,[5],12:i:- were DT20a, U311 and DT193. The same phage types were also found in S. Typhimurium isolates, although with a much lower prevalence. DT7 and DT120 were ascribed to the two serovars at comparable levels. DT104, DT2 and DT99 were ascribed exclusively to S. Typhimurium, and almost all the other phage types identified were more related to the latter serovar. Such data confirm that phage typing can provide an indication of the biphasic or monophasic state of the strains investigated and could therefore support serotyping results. However, phage typing cannot be used as the definitive method to differentiate the two serovars, as part of the phage types were detected for both serovars and, in particular, all phage types found for S. 4,[5],12:i- were found also for S. Typhimurium.