The

<OUTDIR>/summary_tables/*.kofamscan.tsv.gz

file currently contains all hits the tool finds for an ORF. This makes the table quite difficult to use and in all use cases I can think of, having only the best scoring hit is what one wants. I'm therefore asking whether I could just update the script that creates the file so it outputs only the best scoring. The original table as produced by the tool will remain in

<OUTDIR>/kofamscan

. @Danilo Di Leo and others, what do you think?

I did think that there are cases where multiple KO can be valid for one ORF. But I am not familiar with the output file so maybe this is accounted for.

Something like hmmrank, I suppose? I think it's a good idea, it needs to be pointed out what @Daniel Straub is saying though. Probably somewhere in the documentation.

I suppose this is why the tool outputs them. In general I think alternative assignments are to homologs and in almost all cases there's a clear "winner", i.e. best hit. As I mentioned, the original output will always be kept as is and my question applies to the table we produce for easy use (as the others in

summary_tables

). In almost all cases I can think of, multiple entries for ORFs is going to confuse and possibly create errors as it might multiply the number of times an ORF is counted. At first I thought we should output the table as it is plus a version that is subset to only one per ORF, but I thought that might be more confusing for most users. And, sure, this should of course be documented if we choose this way.

I believe the original kofamscan output has a column that the best hit is marked with a star!

Yes, but only when the score is greater than the threshold.