how to calculate the mean sequencing depth of RNA-seq

[wangli]

Dear users

I tried to calculate the mean sequencing depth of RNA-seq. Firstly, I utilized GATK to calculate that with the whole genome as reference. It turned out that some samples with around 15 million mapped reads only have around 2.5* mean sequencing depth. That looks to me weird. Some people suggested me to calculate it by hand, such as: (number of mapped reads * read length)/total length of the species' transcriptome

It looks reasonable. How do you calculate the sequencing depth of RNA-seq?
I will appreciate your comments very much!

[woodydon]

You should calculate the sequencing depth of a gene or an exon. It doesn't make sense to have a number representing the overall sequencing depth of RNA-Seq. DNA and RNA-Seq are different.

[puggie]

You could use e.g. Cufflinks or DEseq to calculate the library depths, at least that would provide you a scaling factor for inter-library comparisons and normalization. I think the "sequencing depth" as per DNA-seq is not comparable in this respect at all.

[Jeremy]

The most informative coverage metric for RNA-seq is the number of genes with x number of reads, e.g. 80% of the known genes had at least 10 reads mapping to them, 95% of the known genes had at least one read map.