Genomic test methods

The Illumina raw reads were analyzed using default parameters of the Trimmomatic platform [52] to trim and crop poor quality reads as well as to remove adaptors followed by de novo assembly using SPAdes assembler [53] on default settings to obtain de novo assembled contigs. The genomes were interrogated for presence of resistance- and virulence-associated genes in de novo assembled contigs using the ResFinder v3.0 and VirulenceFinder v2.0, respectively [54, 55]. Multi-locus sequence typing (MLST) was performed using MLST v1.8 [56] and spa-types were determined using spaTyper v1.0 [57]. The STs were clustered into clonal complexes (CC) according to the S. aureus MLST databases [58].

Single nucleotide polymorphisms (SNP) were identified from the trimmed reads with Snippy v. 4.1 [59] using reference genome Staphylococcus aureus NCTC 8325 ({"type":"entrez-nucleotide","attrs":{"text":"NC_007795","term_id":"88193823","term_text":"NC_007795"}}NC_007795). SNPs located within recombinant regions were filtered out of the core SNP alignment using Gubbins v. 2.3.4 [60]. Some samples (n = 7) were removed automatically due to low coverage or proportion of missing data. The SNP-based phylogenetic tree was built using IQ-TREE v. 1.6.9 [61] with the General Time Reversible model with empirical base frequencies and gamma distribution. One thousand rapid bootstrap replications were performed using the ultrafast bootstrap approximation [62]. Additionally, a dendrogram was constructed to depict the presence or absence of the enterotoxin genes within each genome. All annotations were done using iTOL v. 5 [63].