featureCounts: ERROR: failed to find the gene identifier attribute

Question

[v.berriosfarias]

Hello I made my own gtf file from hmmer results and I used it to calculate abundance of genes from the annotated feature of my gtf file using featureCounts program. The error message that I got is the following:

featureCounts -p -O -T 9 -F gtf -t gene -a demo.gff -o demo_fcounts.txt demo_sorted.bam

ERROR: failed to find the gene identifier attribute in the 9th column of the provided GTF file.

The specified gene identifier attribute is 'gene_id'

An example of attributes included in your GTF annotation is 'gene_id "4_hydroxyacetophenone";E-value "3.5e-192";Score "649.8";Target "k141_179264"'

I have read similar issues published but seems like they do not resolve my problem.

This is the structure of my gtf table:

seqid	Source	feature type	start	end	score	strand	phase	attributes
4_hydroxyacetophenone	hmmsearch	gene	3558	5087	2.6e-180	-	.	gene_id "4_hydroxyacetophenone";E-value "2.6e-180";Score "610.5";Target "k141_165782"
4_hydroxyacetophenone	hmmsearch	gene	151	1617	3.5e-192	+	.	gene_id "4_hydroxyacetophenone";E-value "3.5e-192";Score "649.8";Target "k141_179264"

for my knowledge featureCounts must look at the attributes table and match the feature coordinates (in my case the gene feature type coordinates) with the contig ID (e.g. in the first row: K_141_165782) and it seems to not recognize the gene ID. Is it letting some other attribute ? or instead of providing the gene in the seqid column I should provide the contig ID K_141_165782 (first row example) ?

Thanks for your time.

bests,

Valentín.

Answer 1

[GenoMax]

A GTF file should not be a table unless you mean it in a loose sense. It should be a file with lines containing tabular fields. At a minimum remove the header line if it is actually present. Also include a space after the ; for each attribute. Your GTF file should have reference genome/chromosome names that match the fasta reference you used for alignments/indexes.

Consider using the Simple Annotation Formation (SAF) that featureCounts understands and that you can easily make from your table.

[v.berriosfarias]

yes I provided the column names just for showing it was a gtf file. Thanks for the recommendation of the SAF file. It worked!