Count non-coding genes using featureCounts

Question

[Yizhao]

Hi all,

I try to count aligned reads from RNA-Seq using featureCounts. I used the comprehensive GTF file as annotation downloaded from Ensembl database, which included both protein_coding and non-coding genes annotation. featureCounts output a count file and a summary file as something like we want.

But I found the count file didn't include non-coding gene features of which the annotation was included in my GTF file.

Anybody else notice this problem?

Parameter list I submitted was:

 featureCounts -F 'GTF' -t 'CDS' -g 'gene_id' -Q 20 -d 20 -C -s 0 -T ${ncore} \
              -a ${gtf_file} -o ${featurecount_dir}/${raw_fq_dir}/counts.cds \
              ${star_align_dir}/${raw_fq_dir}/star.sorted.bam

Thanks in advance.

[Yizhao]

I've googled this problem very carefully, but I still have no idea about the reason. Thank in advance.

Answer 1

[Devon Ryan]

If you want to include non-coding genes then you can't count using CDS as the feature type (they lack that by definition). Remove the -g option so you're using the default exon value.

[Nicolas Rosewick]

FYI you should never count reads on CDS, always count on exons for RNA-Seq

[Yizhao]

Thanks for your attention. I 've also tried by submitting "-t exon -g gene_id", which attributes exons to gene-leve meta-feature. Ant the non-coding gene features were not included in the resulting count file either.

[Devon Ryan]

Please post a few lines from your GTF including at least one entry for a non-coding gene.

[Yizhao]

Sorry for the bothering.

[Yizhao]

Great, thanks very much. I've checked the exon-based counting resulting file. And it includes all I want.