How to perform VST on my filtered counts data

Question

[mropri]

Hi guys,

I filtered out my raw count data for my gene list by filtering genes with <5 counts in at least 90% of all samples. Now I have a data frame with genes as rows and counts in 81 samples as columns. I tried running variance stabilizing transformation (vst) on this data frame by using the following code :

varianceStabilizingTransformation(myFilt, blind = FALSE)

However, I am getting this error:

Error in (function (classes, fdef, mtable) : unable to find an inherited method for function ‘estimateSizeFactors’ for signature ‘"data.frame"’

I know there are other posts that asked similar questions and I have tried to understand them but am still confused on why it is not working and giving me this error. Please, any help is appreciated.

[Kevin Blighe]

Cross-posted: https://support.bioconductor.org/p/9144049/

Answer 1

[swbarnes2]

Go read the vignette, and work through it.

Vst doesn't take a data frame as input. It takes a DESeq object.

And yeah, you also have to run estimateSizeFactors on that object first, just like the error message says.