Mismatches in cancer RNA-seq

Question

[IP]

Hi everyone!

I am mapping RNA-seq cancer samples. In the beginning, I did the mapping with tophat default parameters and the reads mapping percentage was around 60%. As I was working with cancer samples, I did the assumtion that my reads will have a high number of mutations and thus, for a read length of 50 bp I have allowed 4 mismatches and my mapping percentage has increase to 80%, while the number of multiple mapping reads remains for both runs similar.

Do you think that this approach is correct?

Thank you!