124 Chapter 7 coreSNP phylogeny was used as true phylogeny and the consistency index was calculated using the multiple sequence alignments fullSNP alignment as input. Relating mutations to CTX resistance To assess whether certain mutations were linked to CTX resistance, all non-plasmidharbouring ampC E. coli isolates were used. CTX resistance was defined using European Committee on Antimicrobial Susceptibility Testing (EUCAST) guideline standards of CTX MIC>2 mg l−1 (The European Committee on Antimicrobial Susceptibility Testing. 2020). CTX MIC results were obtained from our previous study (Coolen et al. 2019). For each nucleotide position on the reference chromosome, the numbers of resistant and sensitive isolates were counted and tested for adenine versus all other nucleotides, thymine versus all other nucleotides, cytosine versus all other nucleotides, and guanine versus all other nucleotides, creating a contingency table and performing a Fisher’s exact test in R 3.6.1 (Mehta and Patel 1983). To correct for multiple testing, P values were adjusted using false discovery rate (FDR) (Benjamini and Hochberg 1995). Selection of genomic positions of interest Genomic positions with potential roles in CTX resistance were identified based on FDR ≤0.05 for CTX and a consistency index of ≤0.05882353. Annotation of mutation positions was obtained by using the genome annotation of the reference chromosome (GenBank accession no. CP046396) and applying snpEff (version 4.3 t) (Cingolani et al. 2012). The EnteroBase core-genome MLST and whole-genome MLST schemes were used to distinguish core and accessory genes (Z. Zhou et al. 2020). Recombination analysis Gubbins version 2.4.1. (settings: -f 30) was used to detect recombination regions with coreSNP alignment and tree as input (Croucher et al. 2015). Pyseer analysis To compare our homoplasy-associated analysis method to an alternative method, we used Pyseer (version 1.3.6) (Lees et al. 2018). In short, variant calling files (VCFs) were obtained from snippy, and bcftools (version 1.11) was used to merge and filter the VCFs from all samples to a single VCF (Li et al. 2009). A phylogenetic distance file was calculated by using the phylogeny_distance.py included in Pyseer on the corrected for recombination phylogeny of Gubbins. Finally, the distance, trait and VCF file was used to run Pyseer (default fixed effects) with settings --min-af 0.01, --max-af 0.99.
RkJQdWJsaXNoZXIy MTk4NDMw