Z-score normalization to normal samples

[antass]

I'm exploring the TCGA Pan-Cancer dataset which includes 32 studies. What samples are used as background when I select expression with normalization to normal samples?

Are there adjacent normal samples available in all of the 32 studies? I'm struggling to find any detailed info on that. In one webinar, the z-score calculation with normals as reference is mentioned, but only in the MSK Prostate study.

Specifically, for my gene of interest, when I look at the Cancer Types Summary tab, the graph shows significant over expression in cholangiocarcinoma (94% of tumor samples have high mRNA status (z-score > 2)). Does this mean over expression of cholangio tumors as compared to

• only cholangio normals?
• only prostate normals?
• all normals in all studies (in which case I'd like to know which studies were included)?

Thank you very much for your help!

[pieterlukasse]

Good question. When I look at the sample type in the pancancer TCGA studies, I see only the following:

So reference normals don't seem to be loaded to the portal.
@ritikakundra any ideas where to find the normals used? Maybe it makes sense to add a link to that as part of the genomic profile description?

[rmadupuri]

Hi @antass sorry for the delayed response! In case you're still interested:

What samples are used as background when I select expression with normalization to normal samples?

The profile shows tumor sample z-scores calculated with reference to expression in normal samples.

Are there adjacent normal samples available in all of the 32 studies? I'm struggling to find any detailed info on that. In one webinar, the z-score calculation with normals as reference is mentioned, but only in the MSK Prostate study.

Normal sample data is available for 16 TCGA Pancancer studies. You can find more details in cBioPortal/datahub#1241

Specifically, for my gene of interest, when I look at the Cancer Types Summary tab, the graph shows significant over expression in cholangiocarcinoma (94% of tumor samples have high mRNA status (z-score > 2)). Does this mean over expression of cholangio tumors as compared to

only cholangio normals?
only prostate normals?
all normals in all studies (in which case I'd like to know which studies were included)?

This indicates over expression of cholangio tumors compared to cholangio normals from the same study.

[rmadupuri]

@pieterlukasse currently only the tumor sample profile with z-scores calculated relative to expression in normal samples is loaded into the portal. We have two additional profiles:

1. mRNA expression of normal samples
2. mRNA expression of normal samples, with z-scores relative to all normal samples

These profiles are not yet loaded into the portal, but you can access this data in the pan_can_atlas studies under the normals folder on Datahub. There was a prior discussion about supporting these profiles, and we should revisit this. CC’ing @jjgao and @ritikakundra - correct me if I’ve missed anything.

We will update the README to include details about this profile.