Schematic overview of the GenTB pipeline. Raw sequence data is quality checked and adapter trimmed before alignment to the H37Rv reference strain (accession AL123456). Variants are called with Pilon, and variant matrices used by the prediction models are prepared using custom scripts available on Github. The analysis will fail if quality criteria are not met (blunt end arrows). Numbers represent the three moments in the pipeline where users can upload their data to predict resistance for their isolate.