Normalize Count Data Intrasamples with limma voom

Question

[Vitor1]

Hi guys,

I'm working with TCGA samples from SKCM dataset and I will look only at tumoral samples and try to fit linear regression models in it, and if possible to later try to validate the regresison model in some melanoma micorarray data.

For what I have read, because I will have only melanoma samples (one group only), I'm wondering if it's correct to normalize the TCGA-SKCM count data with the limma voom function, and consider the patients as "replicates", since I will not have a design matrix.

This is the first time that I will do this and I appreciate any suggestion

Thanks!