Question about mixing TPMs/RPKMs in deconvolution analysis

Question

[marina.wakid]

Hi all,

I have a rather silly question that I'm fairly certain I know the answer to, but since I have never performed cellular deconvolution myself, nor do I have experience in typical workflows for the analysis/using the tools needed, I'll ask anyway:

I'd like to use this shiny app that performs deconvolution for you: https://voineagulab.shinyapps.io/BrainDeconvShiny/, having reference datasets and two algorithms built in makes in convenient to use. The authors ask for input at the RPKM level, since the reference data is in RPKM. Would it be okay to use TPMs as input?

Thanks!