Creating A Gff File With A Nonstandard Attribute

Question

[dolevrahat]

Hello

I am trying to use HTSeq to map counts from an RNASeq assay in order to preform a differential expression analysis on lincRNAs. I have a lisf of the lincRNAs I want to use in the analysis in a BED file, which I converted into a GFFfile using Galaxy's bed_to_gtf_convertor script. An example attribute line from my file:

exon linc|3P|XLOC_000004|TCONS_00000005:2.63081|141AA|61AA|;

The problem is that lincRNAs are not a standard attribute in GFF files attribute columns, so when I attempted run the HTSeq.counts setting idattr to 'linc', I received the following error message:

“Feature linc|3P|XLOC_000004|TCONS_00000005:2.63081|141AA|61AA| does not contain a 'linc' attribute”

since the string 'linc' does appear in many rows of my file's attribute column, I suspect something may be wrong with the attribute strings formatting (wrong delimiters etc.), but am not sure what would be the correct format.

I will appreciate any advice on how to format these strings, or any other ideas on how I can make the attribute column useable

Thanks in advance

Answer 1

[Devon Ryan]

The example you posted contains a "Feature" attribute that happens to contain the word "linc" in it. If the file only contains lincRNAs, then just set the idattr to "Feature".