ROCker models for reliable detection and typing of short-read sequences carrying <i>mcr</i> , <i>erm</i> , <i>mph</i> , and <i>lnu</i> antibiotic resistance genes

Quantitative monitoring of emerging antimicrobial resistance genes (ARGs) using short-read sequences remains challenging due to the high frequency of amino acid functional domains and motifs shared with related but functionally distinct (non-target) proteins. To facilitate ARG monitoring efforts using unassembled short reads, we present novel ROCker models for <i>mcr</i>, <i>mph</i>, <i>erm</i>, and <i>lnu</i> ARG families, as well as models for variants of special public health concern within these families, including <i>mcr-1</i>, <i>mphA</i>, <i>ermB</i>, <i>lnuF, lnuB,</i> and <i>lnuG</i> genes. For this, we curated target gene sequence sets for model training and built these models using the recently updated ROCker V2 pipeline (Gerhardt et al., in review). To validate our models, we simulated reads from the whole genome of ARG-carrying isolates spanning a range of common read lengths and used them to challenge the filtering efficacy of ROCker versus common static filtering approaches, such as similarity searches using BLASTx with various e-value thresholds or hidden Markov models. ROCker models consistently showed F1 scores up to 10× higher (31% higher on average) and lower false-positive (by 30%, on average) and false-negative (by 16%, on average) rates based on 250 bp reads compared to alternative methods. The ROCker models and all related reference materials and data are freely available through http://enve-omics.ce.gatech.edu/rocker/models, further expanding the available model collection previously developed for other genes. Their application to short-read metagenomes, metatranscriptomes, and PCR amplicon data should facilitate more accurate classification and quantification of unassembled short-read sequences for these ARG families and specific genes.IMPORTANCEAntimicrobial resistance gene families encoding <i>erm</i> and <i>mph</i> genes confer resistance to the macrolide class of antimicrobials, which are used to treat a wide range of infections. Similarly, the <i>mcr</i> gene family confers resistance to polymyxin E (colistin), a drug of last resort for many serious drug-resistant bacterial infections, and the <i>lnu</i> gene family confers resistance to lincomycin, which is reserved for patients allergic to penicillin or where bacteria have developed resistance to other antimicrobials. Assessing the prevalence of these genes in clinical or environmental samples and monitoring their spread to new pathogens are thus important for quantifying the associated public health risk. However, detecting these and other resistance genes in short-read sequence data is technically challenging. Our ROCker bioinformatic pipeline achieves reliable detection and typing of broad-range target gene sequences in complex data sets, thus contributing toward solving an important problem in ongoing surveillance efforts of antimicrobial resistance.