IDseq uses a third-party tool called SRST2 (click this link for all of the steps involved in the program) for AMR identification. This tool uses Bowtie2 to align the reads from the raw .fastq file to the ARG-Annot-2 database. Importantly, while SRST2 has a default threshold requiring at least 90% coverage breadth of the reference gene at 5x depth, IDseq requires no threshold for coverage breadth. Therefore it is up to the user to select this threshold by exporting the heatmap and applying a filter in Excel, or a similar program. To filter for higher-confidence AMR genes, you may consider requiring a minimum threshold of 20% coverage. We chose this threshold based on Langelier et al. 2019.