agat_convert_sp_gff2gtf: Is there a way to retain identical isoforms?

Question

[VenGeno]

Hi Jacques and AGAT users,

I am wondering whether there is a way to maintain identical isoforms when we convert? When I use it removes 28 isoforms (identical). It interferes with my Kallisto/IsoformSwitchAnalyzeR pipeline given that it requires 28 identical isoforms removed during the conversion. Thank you in advance.

PS: I asked the same question there at Github. Apologies for crossposting.

[GenoMax]

Why are there identical isoforms? What is their biolofical significance?

[VenGeno]

The original gff3 and fasta files are coming from the potato genome database. I was hesitant to remove those 28 isoforms given they are there in the database. I manually aligned a couple of them just to compare as well. There are no differences and validate the AGAT removal. I guess I need to rerun Kalllisto again to make the Kallisto index without those 28.

Answer 1

[VenGeno]

I am posting Jacques's answer just in case someone wants to remove identical isoforms. Thank you, Jacques!

Currently, it is possible to skip some tests as the identical isoform one but only from the agat_sp_kraken_assess_liftover.pl script. The easiest for you would be to remove

if(! $no_check or   grep( /check_identical_isoforms/, @$no_check_skip ) ) {
    #check identical isoforms
    dual_print ($log, file_text_line({ string => "Check$check_cpt: remove identical isoforms", char => "-", prefix => "\n"}), $verbose );
    _check_identical_isoforms($log, \%omniscient, \%mRNAGeneLink, $verbose);
    dual_print ($log, file_text_line({ string => "   done in ".(time() - $previous_time)." seconds", char => "-" }), $verbose );
    $check_cpt++; $previous_time = time();
}

from the OmniscientI.pm code.

We might try to find an easy way to avoid some tests like that but it will not be in a near future