Is TMM apropriated for Welch test between gender?

Question

[Znow]

Hello, I'm new to bioinformatics and I have a question that's bothering me. I downloaded the data from TCGA and carried out the normalization process using TMM followed by VOOM. I would like to know, can I use the normalized counts to apply the Welch test for mean difference between sex? Or would it be better to use transcript per million, like the ualcan.com website uses?

[ATpoint]

If this is RNA-seq, and you downloaded raw counts which you now normalized by edgeR and processed with limma-voom, then you can directly use limma for a test of differential expression. There is no need for manual testing. It is (together with some other approaches, and for a variety of reasons) the most powerful option.

[Znow]

This is Rna-seq data normalized by edgeR, i used tumor vs normal tissue for this. I just Want to know why platforms like oncodb and ualcan uses welch t test for diference in mean for clinical features like sex, stage and etc. Thais plataforms uses the Transcripts per million as measure, can i use the normalized counts for this ?

Answer 1

[ana]

The best method for this would probably use the raw counts whitout any normalization or treatment