Read-based detection of virulence factors and estimation of E. coli intrapopulation diversity.

Genome and metagenome virulence profiling was examined using the Virulence Factors Database (VFDB; http://www.mgc.ac.cn/VFs/) (64) filtered for E. coli specifically. Metagenomic reads were mapped against the VFDB database, and gene presence or absence was determined by the number of reads recruited by the VF genes (≥1×) and the length of the gene that was covered by reads (≥70%; lower gene abundance or coverage was considered gene absence). E. coli intrapopulation diversity was estimated by calculating the average nucleotide identity of the metagenomic reads (defined as ANIr) mapping to the reference genome (pathotype isolate or reference commensal) with a percent identity between 90 and 100%, i.e., reads representing the total E. coli population in a sample, based on competitive mapping. For this task, the function enve.recplot2.ANIr of the R package enveomics.R v1.3 (65) was used as described previously (https://github.com/lmrodriguezr/enveomics/tree/master/enveomics.R).