For example, see this gene (nad1) in ENA: http://www.ebi.ac.uk/ena/data/view/ABI60879

If you look at the XML for that gene you see the following:

             DQ984518.1: 324706 .. 325091 ,
  complement(DQ984518.1:  24417 ..  24498),
  complement(DQ984518.1:  22828 ..  23019),
             DQ984518.1:   3484 ..   3542 ,
  complement(DQ984518.1: 153702 .. 153960)

Which shows 5 exons joined out of phase and out of order. Is there a valid GTF representation of this?

How to dump a 'non-canonically spliced' gene into GTF? i.e. what's the recommendation?

Cross-posted on biostars

  • $\begingroup$ Maybe you could explain better what is the particularity of splicing for this gene. It is not easy to see in the link you give. $\endgroup$
    – bli
    Jun 21, 2017 at 9:36
  • $\begingroup$ Yes, that's much clearer. I don't have an answer, though $\endgroup$
    – bli
    Jun 21, 2017 at 9:48

1 Answer 1


To my knowledge there's no defined way to deal with that in GTF. GFF3 handles trans-splicing (you'll have to scroll down to "trans-spliced transcript") by giving an individual transcript multiple parents (e.g., ID=some_transspliced_gene;Parent=gene1,gene2). You could use the same methodology with GTF files, but just note that it'll break most downstream programs.

  • 2
    $\begingroup$ To be fair, annotating this in spec-compliant GFF3 would break the vast majority of GFF3 parsers, too. :-) $\endgroup$ Jun 21, 2017 at 18:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.