Hi everyone, We are working with DeepARG and RGI through nf-core Funcscan, screening for AMR genes. We have been doing some general comparison between the two pipelines (RGI and DeepARG) and came across an interesting result. The scatter plot (attached) shows the start position (x-axis) of ARM genes found by DeepARG (blue) and RGI (orange). The question we pose with this result is why DeepARG doesn't seem to find any AMR genes starting after about the 1000 bp mark? Has anyone seen this, or have any ideas as to why DeepARG may only locate genes at the beginning of the contigs? We didn't find any settings to suggest a cutoff, and are investigating whether DeepARG may only classify the first gene it finds and then proceed to the next contig. Any insight would be great! Thanks!

Caveat: I've not run the pipeline on data myself 😅 but I honestly have no idea. From a pipeline perspective, there indeed isn't any value related to this withi nteh pipeline - we only have a length cut off for BGC screening but that's to remove anything less than 1000. Otherwise DeepARG has a number of DIAMOND alignmetns set to 1000 but also doesn't explain it... Do I think it must be something t odo wit htehe tool...