RNA-Seq:Different mapping rates between samples

[stemshell]

Hi everyone,

I have paired end reads from different samples that I aligned to my reference (mouse) using Tophat2. For one sample (and its replicates), I get a 60% mapping rate. However, for another sample (and its replicates), I get a 90% mapping rate. I want to use cuffdiff2 to determine differentially expressed genes between these samples. However, since the mapping rates are different, I was wondering if cuffdiff2 normalizes for differences in mapping rates? Can I still use cuffdiff2 to get differentially expressed genes or do I need to use a different program like DeSeq2 or EdgeR?

Thanks very much in advance for your assistance.

[dpryan]

Cuffdiff2 (as well as DESeq2/edgeR/limma/pretty-much-everything-else) only ever sees the aligned reads, so it has no concept of a mapping rate. I should note that a 60% alignment rate for a mouse is quite low. You might blast a few of the reads to see if you have contamination or something like that. Also, make sure that you quality trimmed things before alignment, as that could be the difference between the samples.