Independent groups normalisation

Question

[elb]

Hi guys, I have a question about the normalization of RNA Seq data. I have 3 types of cells of which I would like to identify the differentially expressed genes. As reported in many papers and tutorials, data have to be normalized together. However, I'm afraid that the normalisation could affect the expression of the genes of one of the three groups for some reasons that I do not know perfectly. Is there a way to make the groups comparable after making a normalisation internal to each cell type?

Thank you in advance

[Kevin Blighe]

Is there a way to make the groups comparable after making a normalisation internal to each cell type?

RPKM / FPKM data is derived from normalisation on a per sample basis; however, this then renders the samples incomparable via statistical analyses (but do not tell this to the 1000s of researchers who, despite this, go ahead and perform differential expression analysis with RPKM / FPKM data).

[h.mon]

Before trying to solve a problem, check if there is a problem. If you don't know beforehand if the normalization will work or not, why don't you first try and evaluate the results? Examine the diagnostic plots, the number and pattern of differentially expressed genes between each contrast, and so on.