Antimicrobial resistance (AMR) is a recognised global threat with substantial predicted impact on lives, agriculture, and the economy. Metagenomic sequencing is being increasingly used for AMR surveillance and detection, given its capacity for community-level AMR profiling with high-level resolution. This technology has seen an explosion of surveillance efforts and data generation; however, the variation between workflows has direct implications on the sequencing results and their interpretation. In this Personal View, we summarise aspects of the sequencing workflow that need to be considered during metagenomic study design, for meaningful and reliable population-based surveillance. We reflect on the vital role of standardisation for capturing the ground truth of AMR and data comparability and reproducibility, and in addition, review the limitations of the various phenotypic and genotypic methods of AMR detection. We further highlight complex mechanisms of resistance to antimicrobials that could hinder our ability to confidently assess the true AMR burden in the environment and those that are often overlooked during surveillance.
Hughes et al. (2026) studied this question.