How does read length affect sequencing depth for paired end RNA-Seq?

[bob-loblaw]

I'm just wondering if an experiment run using say 50bp reads and that requires a sequencing depth of 200 million reads to fully assess gene expression, what affect will using 100bp reads have on that sequencing depth? Obviously less sequencing depth would be required, but I was wondering if anyone could give me a rough estimate of how much?

[kmcarr]

Originally Posted by bob-loblaw

I'm just wondering if an experiment run using say 50bp reads and that requires a sequencing depth of 200 million reads to fully assess gene expression, what affect will using 100bp reads have on that sequencing depth? Obviously less sequencing depth would be required, but I was wondering if anyone could give me a rough estimate of how much?

When using RNA-Seq to study gene expression read length is not a significant factor; what matters is read counts. Whether you align 100bp paired reads or a 50bp single read to a gene they each still only count as one read. It is counts that give you analytical power in expression studies, not how long your alignments are.

[kcchan]

Originally Posted by kmcarr

When using RNA-Seq to study gene expression read length is not a significant factor; what matters is read counts. Whether you align 100bp paired reads or a 50bp single read to a gene they each still only count as one read. It is counts that give you analytical power in expression studies, not how long your alignments are.

What kmcarr is saying is correct; counts mean more than the read length in expression analysis. However, a paired end read 2x50bp will be more beneficial than a 100bp single read as it enables the ability to remove PCR duplicates with more precision.