2.4. Whole genome sequence (WGS) analysis
For selected ESBL-producing E. coli strain, genomic DNA was extracted using a PureLinkTM Quick Gel Extraction Kit (Life Technologies, Carlsbad, CA), and a genomic paired-end library (75 x 2 bp) was prepared using a Nextera XT DNA Library Preparation Kit (Illumina Inc., Cambridge, UK) according to the manufacturer’s instructions. The whole genome was sequenced on the NextSeq platform (Illumina). De novo genome assembly and contig annotation was carried out using CLC Genomics Workbench 12.0.3. Multilocus sequence type (MLST), plasmid replicons, resistome and serotype were identified using MLST v2.0 (Larsen et al. , 2012), PlasmidFinder v2.1 (Carattoli et al. , 2014), ResFinder v3.2 (Zankari et al. , 2012), and SerotypeFinder v2.0 (Jenkins, 2015) tools, respectively, from Center for Genomic Epidemiology (http://www.genomicepidemiology.org/). Clinically important virulence factors were detected and compared by ABRicate v0.9.8 (https://github.com/tseemann/abricate) using data from the Escherichia coli Virulence Factors (https://github.com/phac-nml/ecoli_vf) and the Virulence Factor Database (VFDB) (http://www.mgc.ac.cn/VFs/). Heavy metal (HM) and biocides genes were detected using the BacMet2 experimentally confirmed database (http://bacmet.biomedicine.gu.se). For whole genome of selected ESBL-producing E. coli identified in this study, a minimum spanning tree was constructed in Enterobase using the MSTree V2 algorithm and the wgMLST scheme (https://enterobase.warwick.ac.uk/species/index/ecoli). This scheme consists of 25,002 pan-genome genes present in E. coli genomes, which represented most of the diversity in Enterobase at the time (March 2020) (https://bitbucket.org/enterobase/enterobase-web/wiki/Escherichia%20Statistics). All images were generated with iTOL v.5.5 (https://itol.embl.de).