RNA-Seq data from FFPE samples

[Payton Yau]

Hi all, 

I have been trying to align RNA-Seq data from human FFPE (formalin-fixed specimen) samples. I used Rsubread default setting for the alignment and it came out about 16% ~ 30% of successfully alignment rates,  it is (relatively) low, but someone said it is "normal" (https://www.biostars.org/p/339745).

I am wondering, does any one have the experience in processing degraded samples by fine tuning the align function?

Thank you for your help.

Best Regards,

Payton

[Wei Shi]

Dear Payton,

I dont have much experience with FFPE samples but you should take caution if you would like to change the mapping parameters. You can easily get a lot more reads mapped by relaxing mapping thresholds but a lot of them might be incorrect mappings. I would rather to have a small set of reads mapped with good confidence for further analysis instead of having a large set of reads mapped with a low confidence.

Regards,

Wei