An improved DNA array-based classification method for the identification of Salmonella serotypes shows high concordance between traditional and genotypic testing

• Published in: PloS one 2018-12, Vol.13 (12), p.e0207550-e0207550
• Main Authors: Robertson, James, Yoshida, Catherine, Gurnik, Simone, McGrogan, Madison, Davis, Kristin, Arya, Gitanjali, Murphy, Stephanie A, Nichani, Anil, Nash, John H E
• Format: Article
• Language: English
• Subjects: Analysis; Antigens; Antimicrobial agents; Arrays; Bacterial genetics; Biological markers; Biology and Life Sciences; Biomarkers; Classification; Classifiers; Deoxyribonucleic acid; DNA; DNA microarrays; DNA probes; Drug resistance; Food Safety; Gene amplification; Genetic aspects; Genetic testing; Genomes; Genomics; Genotyping Techniques; Identification; Internet; Laboratories; Limit of Detection; Medicine and Health Sciences; Methods; Oligonucleotide Array Sequence Analysis - methods; Oligonucleotides; Public health; Repositories; Research and Analysis Methods; Salmonella; Salmonella - classification; Salmonella - genetics; Serotypes; Serotyping; Serotyping - methods
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0207550

Previously we developed and tested the Salmonella GenoSerotyping Array (SGSA), which utilized oligonucleotide probes for O- and H- antigen biomarkers to perform accurate molecular serotyping of 57 Salmonella serotypes. Here we describe the development and validation of the ISO 17025 accredited second version of the SGSA (SGSA v. 2) with reliable and unambiguous molecular serotyping results for 112 serotypes of Salmonella which were verified both in silico and in vitro. Improvements included an expansion of the probe sets along with a new classifier tool for prediction of individual antigens and overall serotype from the array probe intensity results. The array classifier and probe sequences were validated in silico to high concordance using 36,153 draft genomes of diverse Salmonella serotypes assembled from public repositories. We obtained correct and unambiguous serotype assignments for 31,924 (88.30%) of the tested samples and a further 3,916 (10.83%) had fully concordant antigen predictions but could not be assigned to a single serotype. The SGSA v. 2 can directly use bacterial colonies with a limit of detection of 860 CFU/mL or purified DNA template at a concentration of 1.0 x 10-1 ng/μl. The SGSA v. 2 was also validated in the wet laboratory and certified using panel of 406 samples representing 185 different serotypes with correct antigen and serotype determinations for 60.89% of the panel and 18.31% correctly identified but an ambiguous overall serotype determination.