Research Square
A project-selection audit of public urinary-source Escherichia coli genomes from China distinguishes purpose-enriched NDM and mcr signals from CTX-M recurrence across projects
By Yu B, Wang C, Wang Z and 1 others.
Abstract
Abstract Background Public bacterial genome archives combine surveillance collections, resistance-focused studies, selected clinical subsets, and individual genome reports. The reasons for sequencing can therefore affect pooled resistance proportions. This study traced project-level selection mechanisms in public urinary-source Escherichia coli genomes from China and assessed clonal and mobile resistance signals. Results The cohort contained 512 genomes from 29 BioProjects. Original publications were identified for 19 projects that contributed 362 genomes. BioProject or case records supported the remaining 10 projects and 150 genomes. Eighteen resistance-targeted or enriched projects supplied 350 genomes (68.4%). All 144 NDM-positive genomes and 142 of 145 mcr-positive genomes arose from this stratum. CTX-M occurred in 25 of 29 projects. It was present in 44/92 genomes (47.8%) from a syndrome-defined clinical collection and 32/55 (58.2%) from two representative clinical subsets. Its pooled proportion varied by no more than 4.2 percentage points in leave-one-project-out analyses. Major Achtman lineages showed distinct resistance profiles. ST410 and ST167 were NDM-rich. ST131 and ST1193 mainly carried fluoroquinolone resistance-associated and CTX-M genotypes. A recombination-filtered ST167 phylogeny recovered several related NDM-positive sublineages. Local sequence contexts associated NDM mainly with IncX3-like regions and mcr mainly with IncI2- and IncX4-like regions. Recurrent integron cassette modules occurred across multiple projects and lineages. Conclusions The project-selection audit separated concentrated NDM and mcr signals from CTX-M recurrence across projects. Source-study audits should accompany accession-level analysis of heterogeneous public genome collections. Together, these analyses estimate archive composition while retaining clonal and mobile-element context.



