AMRProfiler: A Comprehensive Tool for Identifying Antimicrobial Resistance Genes and Mutations Across Species

Introduction
Antimicrobial resistance (AMR) is a critical global health challenge. The advent of affordable whole genome sequencing has facilitated in silico methods for identifying AMR gene content, detecting known resistance mechanisms, and exploring novel mechanisms. Accurate detection of resistant bacterial isolates and antimicrobial resistance genes (ARGs) is vital for understanding AMR evolution, emergence, and optimizing treatments. Primary AMR data repositories include ResFinder, the Reference Gene Catalog, and CARD. We introduce AMRProfiler, a bioinformatic tool that integrates these databases to simplify the identification of AMR genes, point mutations in core genes, and mutations in rRNA genes.

Methods
AMRProfiler integrates data from ResFinder, the Reference Gene Catalog, and CARD, filtering redundant records to yield 7,600 unique AMR gene entries. For point mutations, 266 core genes with 4,380 mutation records were compiled across the databases. rRNA gene mutations were analyzed using 18,640 reference genomes from RefSeq, recording the number and type of rRNA gene copies in each genome. Users can query AMR genes by setting thresholds for identity, coverage, and protein start site. The tool also allows point mutation searches in core genes based on reference assemblies, identifying mutations across 18,640 species and uniquely reporting nearby mutations with known AMR potential. Furthermore, it facilitates point mutation searches in rRNA gene copies within input assemblies.

Results
AMRProfiler was validated using Acinetobacter baumannii assembly GCF_033142795.1 alongside AMRFinder, ResFinder, and CARD. For AMR genes, AMRProfiler detected all genes identified by the other tools and uniquely identified the genes adeC, amvA, dfrA40, and abaF. For point mutations, AMRFinder detected only one point mutation in gyrA, whereas AMRProfiler identified multiple point mutations in 27 genes, including experimentally validated mutations in gyrA, gyrB, parC, and parE. Additionally, AMRProfiler identified all rRNA gene copies and their mutations.

Conclusion
AMRProfiler enhances precision in identifying AMR genes, linking AMR genotypes and phenotypes. Its robust and user-friendly design supports comprehensive AMR gene, core gene mutation, and rRNA mutation analysis across multiple species. An online web server for AMRProfiler will be launched in the coming months.