gffread query for gff to gtf conversion

Question

[prasundutta87]

Hi,

I was just wondering if one can explain to me what is happening with the tool gffread while converting one gff file to gtf file. I used the following command:

gffread <GFF3_file> -T -o <GTF_file>

When I checked the unique chromosomes/scaffold names to check if the conversion was correct, both gff file and gtf file had different numbers. On digging, I found that some scaffolds (mentioned below) were present exclusively in gff and were not converted to gtf. Example:

NW_020228737.1  RefSeq  region  1       24009   .       +       .       ID=id827864;Dbxref=taxon:89462;Name=Unknown;breed=Mediterranean;chromosome=Unknown;dev-stage=adult;gbkey=Src;genome=genomic;mol_type=genomic DNA;sex=female;tissue-ty
##sequence-region NW_020228738.1 1 24016
##species https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=89462
NW_020228738.1  RefSeq  region  1       24016   .       +       .       ID=id827865;Dbxref=taxon:89462;Name=Unknown;breed=Mediterranean;chromosome=Unknown;dev-stage=adult;gbkey=Src;genome=genomic;mol_type=genomic DNA;sex=female;tissue-ty
##sequence-region NW_020228739.1 1 23921

They are unplaced scaffolds, but why were they not converted? Can someone let me know if I am missing something?

[prasundutta87]

I have posted this question in the gffread github page just now. Hopefully, I may get a reply there..if I get it soon, I will post the reply here as well..

Answer 1

[prasundutta87]

Got this reply..

GTF is a transcript-oriented format, it only supports a very limited set of features (as opposed to GFF3), namely only transcript-describing features (exon, CDS and that's pretty much it). The "region" feature here is not something that GTF cares about -- and frankly neither does gffread (which by default only cares about transcript features as well), since it's not really a transcript or annotation feature, it's a rather useless acknowledgment of a genomic scaffold's presence -- and there is no GTF version of it.