htseq-count shows fatal error every time



Hello everyone.

I am trying to operate htseq-count in galaxy. I have used HISAT2 BAM files (I am using Arabidopsis WT and Double Knock Out) and TAIR10.GFF3 annotation file.

But every time it shows fatal error after a while.

Requesting you all to provide your invaluable suggestions.

Thanking you,



Try using the gtf version of the annotation instead. This tool does not accept gff3 formatted annotation.

This can get a bit confusing since the tool form states gff as an input, and will accept files with that datatype assigned. This is because many data sources provide gff+gft hybrid data – and the only difference between those two is the last column (attributes). Some also release gff3 data that is given a gff datatype (including this source: The gff3 format is structured and labeled much differently than gff or gtf.

These data are incredibly difficult to automatically assign a datatype to, because there can be so much variation from sources. Custom header lines can interfere, blank lines, etc. It is usually a good idea to check the format yourself once in Galaxy to confirm the actual type. Sometimes headers need to be removed in gff or gtf datasets before certain tools will consume them properly.

This FAQ explains each, how they are similar/different, and reformatting/transformation tips.


Hello Jennifer,

Thanks for your reply and explanation.

As per your suggestion, I have used gtf file and it seems like working (still running, didn’t show any fatal error after a while). Let’s see what comes at the end.

Thanks again.