RNA-Seq requirements for both gene expression quantification and variant calling

Question

[Ozzie]

I would like to use RNA-Seq on fresh tumor tissue for both gene expression quantification and variant calling.

Do I need stranded RNA information?

Do I need polyA enrichment based or rRNA depletion based rRNA removal?

How many data (sequencing reads or coverage) per sample do I need?

Thanx!

[cpad0112]

Being curious, would like to know the basis of your sample size.

Answer 1

[Devon Ryan]

1. No, but it's not much more expensive and you get a bit nicer data.
2. You can do either, but poly-A enrichment would probably make the most sense.
3. As much as you can afford if you want to do variant calling. Honestly, I'd do 15-20 million reads for RNAseq and then pair that with exome sequencing. Otherwise you're going to have to really up your depth to get decent coverage of lowly-expressed genes and will still be left with the problem of allele-specific expression.