Why there are no reads for gene quantification but many reads can be quantified for exon using featureCounts

Question

[mina95921m]

Hi, I used featureCounts to quantitate whole genebody and exon-only region of the same gene using the same bam files, and unexpectedly found that exon-only region successfully assigned 4000k fragments while 0 fragments assigned whole genebody region, so I wonder why did this happen and how to quantify gene body region correctly?

featureCounts -T 10 -p -C -B -s 2 -a Mus_musculus.GRCm38.99.gtf -o exon.quant 19R774.sort.bam featureCounts -T 10 -p -C -B -s 2 -t gene -a Mus_musculus.GRCm38.99.gtf -o gene.quant 19R774.sort.bam

Answer 1

[lieven.sterck]

can you provide a bit more data/result to back up what you are saying? Is it for all genes/exons in the gtf file or only for part of them?

give some examples for instance for genes in which both approaches differ.

[mina95921m]

only part of them, and the gene for example:

grep ENSMUSG00000058207 gene.quant

ENSMUSG00000058207 chr12 104338486 104346144 + 7659 0

grep ENSMUSG00000058207 exon.quant

ENSMUSG00000058207 12;12;12;12;12 104338486;104340494;104342939;104344127;104345233 104338545;104341150;104343212;104344280;104346144 +;+;+;+;+ 2057 433393

[lieven.sterck]

hmm, that does not look quite correct.

just as with the gene approach you should get one line per exon. Do double check the formatting of the output (input?)

[mina95921m]

I'm sorry I didn't describe it clearly. Genequant=exon+intron region , exonquant means all exon region of the gene, so the result of exon.quant also one line per gene including all the exon region. Actually, I want to get reads abundance of intron region using genequant-exonquant.

using IGV to check ENSMUSG00000058207 reads abundance find high abundance in exon region which means the result of exon.quant is right while something wrong in gene.quant.

[lieven.sterck]

not sure I follow, but you can't get counts over two features? you can get counts for introns, for exons, for genes, but not for exon + intron features. These tools work on a feature mentioned in the third '?) column of you gff/gtf file

[mina95921m]

I use -t gene to count reads covering the full length of the gene which both include intron and exon region. so exon+intron =gene, one feature ,I don't know if my understanding is correct?

[lieven.sterck]

yes, that is correct. (I was temporarily under the impression you wanted to count on exon + intron feature, rather than gene as a feature)

[lieven.sterck]

I do notice that in the gene.quant you mention above it says 'chr12' and in the exon one is says '12' . can you verify this (and also make sure you are consistent in this for all files you use, thus all 12 or all chr12 )