Entering edit mode
8.2 years ago
mmacd
▴
20
Hi all,
I have been looking at different gff3 to gtf converters, but cannot find a good one that works well for gff3 files downloaded from NCBI Refseq assemblies. I am trying to compare (using the program Eval which only takes in gtf files) an existing refseq annotation with one I created using Maker. Maker provides a script to convert its gff3 output to gtf specifically so that Eval can read it in. This worked great on my Maker output, but did not work for the RefSeq gff file.
I have also used Cufflinks gffread tool with no luck either.
Any help is greatly appreciated!
What's the genome and why do you say the software you found doesn't work well?
The genome is the Chinese hamster. When I run Maker's converter script, the gtf output file only contains CDS features and files is about 10 fold smaller.
That is what the GTF is supposed to contain, not all of the alignments and other features (the acronym "GTF" and the specification explains these things).
Can you send me the link to the documentation for GTF for where it says its supposed to only contain CDS features? I have read elsewhere that it describes gene structure and have seen documentation (http://mblab.wustl.edu/GTF22.html) showing GTF's containing exons, utrs etc.
The answer to your question can be found in the document you linked to for the GTF format. The specification for the feature:
GTF is basically GFF, but as the name suggests, is more specific to the features it describes and must have the same nomenclature for the gene IDs and transcripts (unlike GFF which is flexible in this respect).
If my GFF file had "exon" features, shouldn't the GTF? I am looking for a tool that can do the conversion without losing exon information.
If multiple tools do not work that usually means there is something wrong with the input. It would help to link to the NCBI page so people could see the file. Also, try to link to the software you mention because it is not always clear what script/tool is being discussed. EDIT: Disregard this, there does not appear to be an issue. OP was not clear what a GTF is supposed to contain.