PMC

Flow chart of isolate extraction, sequencing, and data analysis.

Proportions of antimicrobial resistance genes associated with isolates in each phase of study. The proportions of isolate genomes containing predicted antibiotic resistance in each subject and time period are presented as a heat map. Protein sequences known to confer E. coli antibiotic resistance were queried against the translated isolate genomes using CARD-RGI. The total number of resistant protein sequences present is presented as a proportion of the total number of isolates in that time period (in parentheses), with multiple copies in some isolates. The resistance sequences are listed by the category of antibiotic to which they provide resistance. Those sequences marked with an asterisk (*) provided antibiotic resistance due to a sequence mutant or variation (detailed in ).

Virulence gene presence and absence. The proportions of isolate genomes containing virulence factors in each subject and time period are presented as a heat map. Gene sequences of known E. coli virulence genes were queried against the isolate genomes using LS-BSR. Genes with a significant sequence match (BSR > 0.8) were deemed present, and the total number of present virulence genes is presented as a proportion of the total number of isolates in that time period (in parentheses). Genes involved in ETEC toxin (LT and ST) production are highlighted in bold. Other listed genes are involved in ETEC adhesion or virulence in other E. coli pathotypes, with details in . BSR assay results from individual isolates for each investigated virulence factor gene are presented in .

Phylogenetic analysis of isolates by subject. The whole-genome sequences of the isolates from each subject were compared with 32 previously sequenced E. coli and Shigella genomes (listed in ) using a single nucleotide polymorphism (SNP)-based approach as previously described (, ). SNPs were detected relative to the completed genome sequence of laboratory isolate E. coli UTI89 using the in silico Genotyper (ISG) tool (). A range of 171,581 to 175,765 conserved SNP sites which were present in all of the genomes analyzed were concatenated into a representative sequence for each genome. A maximum-likelihood phylogeny with 100 bootstrap replicates was inferred using RAxML v.7.2.8 (). Phylogenetic trees of all isolates in each subject are listed in order of increasing disease severity. Isolates collected prechallenge are shown in yellow, isolates collected during challenge in red, isolates collected postchallenge in blue, reference strains in gray, and the challenge strain in black with an arrow.

Phylogenetic analysis based on phase of challenge. The whole-genome sequences of the isolates from each subject based on the phase of the challenge were compared with 32 previously sequenced E. coli and Shigella genomes listed in using a single nucleotide polymorphism (SNP)-based approach as previously described (, ). SNPs were detected relative to the completed genome sequence of laboratory isolate E. coli UTI89 using the in silico Genotyper (ISG) tool (). A range of 67,495 to 140,098 conserved SNP sites which were present in all of the genomes analyzed were concatenated into a representative sequence for each genome. A maximum-likelihood phylogeny with 100 bootstrap replicates was inferred using RAxML v.7.2.8 (). Phylogenetic trees of isolates from the three time periods (prechallenge [A], during challenge [B], and postchallenge [C]) show isolates from subject 001 in green, 006 in blue, 008 in orange, 009 in purple, 015 in pink, 016 in red, 004 in yellow, 019 in light blue, reference strains in gray, and the challenge strain in black with an arrow. Trees for individual time points are presented in .