Is additional normalization needed when compare counts between two different projects, like TCGA and GTEx?

Question

Is additional normalization needed when compare counts between two different projects, like TCGA and GTEx?

Closed this issue 6 years ago · 1 comments

I would like to compare the distribution of gene expression between samples from TCGA and GTEx. So I download the RSE from TCGA and GTEx and do scale_counts and combine two data set together for histogram plots. So after combing the two data set, do I need to do additional normalization for comparison?
Thanks

Answer 1 · 2019-06-18T17:29:06.000Z

Hi,

Sorry that I missed this question a while back. In any case, the answer is complicated as normalization across projects is an active area of research. @kasperdanielhansen and others might have more detailed information for you.

As a starting point, check Supplementary File 3 from https://www.biorxiv.org/content/10.1101/618025v1 written by @ShanEllis.

Best,
Leonardo