Key points are not available for this paper at this time.
Abstract The accelerating spread of extended-spectrum β-lactamase (ESBL)-mediated resistance threatens the clinical utility of β-lactam antibiotics across human, veterinary, and environmental health sectors, underscoring the need for reliable genomic frameworks within a One Health context. Public genomic repositories are central to antimicrobial resistance (AMR) research, yet their utility is constrained by annotation inconsistency, nomenclature ambiguity, and biased (non-random) data deposition. Here, we performed phylogeny-guided curation and variant-level fingerprinting of ESBL gene families using 2746 β-lactamase sequences retrieved from UniProt, GenBank, Kyoto Encyclopedia of Genes and Genomes (KEGG), and Comprehensive Antibiotic Resistance Database (CARD), comprising bla CTX-M ( n = 1316), bla TEM ( n = 895), and bla SHV ( n = 535). Conserved motif analysis, domain architecture assessment, and KEGG/CARD cross-validation revealed extensive annotation discordance: only 42.9% (1177/2746) of entries were correctly annotated, 3.8% (104/2746) required reclassification, and 53.3% (1465/2746) remained insufficiently resolved. Phylogenetic reconstruction enabled reclassification of cryptic variants and delineation of species-associated phylogroups consistent with lineage compatibility and horizontal gene transfer. Significant gene-host structuring was observed across species (χ 2 = 479.64, df = 28, P = 1.04 × 10⁻ 83 ), with Escherichia coli enriched for bla CTX-M-15 ( n = 124), Klebsiella spp. for bla SHV-11 ( n = 105), and Salmonella spp. for bla TEM-1 ( n = 52). These dominant allele patterns were independently supported by isolate-level genome surveillance metadata from NCBI Pathogen Detection (Create date ≤ 2022), confirming bla CTX-M-15 predominance in E. coli / Shigella and bla SHV-11 predominance in Klebsiella pneumoniae . Subvariant diversity (Shannon index) was highest in Acinetobacter spp. (H′ = 3.578), Enterobacter spp. (H′ = 3.385), and Proteus spp. (H′ = 3.251), whereas Vibrio spp. exhibited restricted diversity (H′ = 0.95). Morphological stratification confirmed the predominance of Gram-negative bacilli among ESBL-encoding organisms (99%; χ 2 = 33.62, df = 2, P = 5.00 × 10 −8 ). Analysis of database submissions spanning 1998–2022 demonstrated statistically significant temporal shifts in ESBL family representation across years (χ 2 = 254.20, df = 36, P = 1.21 × 10⁻ 34 ), with a marked rise in bla CTX-M representation after 2016 period. Ecological attribution using curated metadata indicated that 85% ( n = 2324) of ESBL entries were linked to species occurring at the human–animal interface, likely reflecting database deposition bias rather than true reservoir prevalence. Collectively, this study provides a high-confidence, variant-resolved ESBL fingerprint derived from curated public databases, highlights systematic annotation limitations, and supports database-aware interpretation of comparative AMR research within a One Health framework.
Hassan et al. (Wed,) studied this question.