Gff file error in HTSeq-count

Question

[y.gladbach]

Hi there, I need your help:). I guess my gff file is not correctly formatted, but not sure. Help is much appriciated. So please let me know, if this is true. I get the an error by calling htseq-count:

samtools view -h {} |  htseq-count -m intersection-nonempty -i Name -t ID -s no - piRNA_GRCh37.gff > {.}.count

The error:

Error occured when processing GFF file (line 1 of file /home/gladbach/Documents/genomes/GRCh37/DASHR/piRNA_GRCh37.gff):
Failure parsing GFF attribute line
[Exception type: ValueError, raised in __init__.py:164]

And my gff file:

chr1    ncbi    piRNA   74625   74655   .   +   .   name=piR-61514;gi;108090112;gb;DQ595402.1;ID=piR-61514
chr1    ncbi    piRNA   75435   75464   .   +   .   name=piR-37026;gi;108096889;gb;DQ598960.1;ID=piR-37026
chr1    ncbi    piRNA   135736  135765  .   -   .   name=piR-43123;gi;108056654;gb;DQ575011.1;ID=piR-43123
chr1    ncbi    piRNA   137313  137343  .   -   .   name=piR-43675;gi;108057724;gb;DQ575563.1;ID=piR-43675

[Devon Ryan]

Where did you get that GFF file? It's highly malformed.

[y.gladbach]

I got it from here: dashr

It's for piRNA

[Devon Ryan]

You might be able to salvage that with sed -e "s/;gi;/;gi=/g; s/;gb;/:gb=/g; s/\;/\; /g" input.gff > modified.gff.

[cmdcolin]

maybe this format would be applicable for the column 9 "name=piR-43675,gi:108057724,gb:DQ575563.1;ID=piR-43675" things are allowed to have multiple values in GFF, it should be comma separated though. The semicolon is reserved as a separator between values. technically name is an official reserved keyword in GFF too so should be "Name=" also

[y.gladbach]

I downloaded now from pirnabank and formatted my file like this (since I trust this website more), would this be correct for gff3 (tab seperated or space? and do I need a header starting with #?):

1       .    piRNA  4493  4520       .  +       . ID=hsa_piR_013426;accession_number=DQ588205
1       .    piRNA  8399  8426       .  +       . ID=hsa_piR_005239;accession_number=DQ577218
1       .    piRNA 16669 16699       .  -       . ID=hsa_piR_016792;accession_number=DQ593109
1       .    piRNA 21997 22023       .  -       . ID=hsa_piR_019669;accession_number=DQ596983
1       .    piRNA 29543 29573       .  -       . ID=hsa_piR_014636;accession_number=DQ590030
1       .    piRNA 41516 41544       .  -       . ID=hsa_piR_016271;accession_number=DQ592181

[Devon Ryan]

That looks better and will probably work (perhaps you'll need a space after the semi-colon).