How does cBioPortal normalize data

Question

[Wenhu_Cao]

I have a dumb question, maybe. I am looking into the cbioportal website for cancer genomic data, it said they don't include normal tissue data.

Then, I wonder how do they get the data, like RNA-seq, CNA, etc. What are they using to compare with the tumor sample data? What is the normalization standard? What is the control group?

[Kevin Blighe]

I edited the tags of your question to include 'cbioportal', so that Cyriac may pick it up.

As far as I know, cbioportal is just stating that it doesn't make data from normals available for download. This does not imply that normals were not used for comparisons during data processing. When you download mutation data, for example, it has already been processed and what you're downloading are somatic variants from a tumour versus normal comparison.

For some of the other data-types, including copy number profiling and RNA-seq, there was not enough funding (or physical biopsy material) available in the original TCGA study such that they could profile all normal samples. For RNA-seq, for example, in a dataset of 500 cancer samples, there may only be 20 matched normal RNA-seq samples available.

[Wenhu_Cao]

Thanks for your interpretation. I understand now:)

[GenoMax]

Do people not look at FAQ sections of portal websites any more?

[Wenhu_Cao]

Actually I read, but I may need to dig deeper than I thought~~

[GenoMax]

While someone from cBioPortal may participate here they may not do so regularly. Your best bet may be to contact them (or the support email) directly with your questions.