From targeted to broad detection: How probe-based hybrid capture enhances antibiotic resistance detection in wastewater

Wastewater-based surveillance has emerged as a promising approach for population-level monitoring of antimicrobial resistance genes (ARGs) beyond clinical settings. Although digital PCR and droplet digital PCR (ddPCR) enable sensitive detection and quantification of predefined ARGs, their limited multiplexing restricts the number of targets analysed in a single assay. In contrast, shotgun metagenomics offers broad resistome coverage but may have limited sensitivity for low-abundance targets in complex matrices. Probe-based hybrid capture combines targeted enrichment of thousands of ARGs with high-throughput sequencing and may enhance sensitivity while retaining broad coverage and sequence-level resolution. However, its performance relative to ddPCR and shotgun metagenomics has not yet been evaluated using the same wastewater samples. Therefore, we compared ddPCR, shotgun metagenomics and probe-based hybrid capture using eight influent samples from one wastewater treatment plant. Samples were selected from a longitudinal dataset to cover contrasting ARG abundance profiles. Eight predefined targets were quantified by ddPCR to compare target detection and relative abundance patterns. Shotgun metagenomics described the resistome but detected intermediate- and low-abundance ARGs less consistently. Under the conditions tested, probe-based hybrid capture detected more ARGs per sample and improved recovery of targets poorly represented in the shotgun metagenomic data. Probe-based hybrid capture abundances were positively correlated with ddPCR concentrations for some targets. These results indicate concordance in relative abundance patterns for selected targets, rather than agreement in absolute concentrations. Probe-based hybrid capture may complement ddPCR and shotgun metagenomics for wastewater AMR surveillance when both broad ARG coverage and sensitive detection are required. However, the findings are based on a small, deliberately selected sample set from a single wastewater treatment plant. Further evaluation using larger, independently selected datasets from different treatment plants, geographical settings and wastewater matrices is needed to determine the robustness and applicability of the approach for routine surveillance.