Hi,

I performed an analysis using DRAGEN RNA pipeline on illumina platform. I have the following output files: quant.sf quant.gene.sf I was wondering which one I have to use to perform a PCA using R and why?

Thank you in advance! Ilaria

Hi,

If you have quant.sf files, then these were produced by the Salmon pseudo-aligner. These likely contain estimated counts for genes. You will first have to process these quant.sf files via tximport and DESeq2, and then run PCA on the variance-stabilised expression levels that are produced via DESeq2.

Some helpful links for you:

1. Transcript abundance files and tximport / tximeta
2. Extracting transformed values
3. Principal component plot of the samples

There is also my own dedicated PCA package, PCAtools, which has a tutorial for performing PCA via the DESeq2 route:

• PCAtools: everything Principal Component Analysis

Kind regards,

Kevin