TCGA clinical feature meaning of "pct_tumor_invasion" for TCGA-UCEC project (Uterine Corpus Endometrial Carcinoma)

Question

[Fdota]

Hello,

I have searched the Data Dictionary Viewer from https://docs.gdc.cancer.gov/Data_Dictionary/viewer/, but I didn't find the description of the feature "pct_tumor_invasion" showing up in the "clinical patient" data for project TCGA-UCEC, does anyone know what this feature means? Thanks!!!!

Answer 1

[Kevin Blighe]

• Edit 24th September 2018: Yet another answer, with further information, A: TCGA patient variable data dictionary
• Edit October 3, 2020: I eventually made a tutorial about this: NIH CDE browser for TCGA clinical data CDE codes

########

If you want to search for one of those terms used by the TCGA, you should go by the CDE code assigned to each term, which should be listed in the third row of the clinical data that you downloaded. For pct_tumor_invasion in the endometrial cancer dataset that I have, the CDE ID is CDE ID 3104403.

Here is where you can search for these: https://cdebrowser.nci.nih.gov/cdebrowserClient/cdeBrowser.html (if you use the CDE ID, search under the Public ID Search tab).

Kevin

[Fdota]

Thanks Kevin, I found the description of the term. May I ask where can I find those CDE ID? I am using GDCprepare_clinic function from "TCGAbiolinks" package by Bioconductor to get those datasets. That pct_tumor_invasion feature is in the "patient" dataset, if you know what dataset the CDE ID is in, like "drugs", "follow_ups", "radiations" or "new_tumor_events"? Thank you!!!

[Kevin Blighe]

No problem! I have never used TCGAbiolinks as I normally download the data myself from the GDC Data Portal using manifest files. However, the TCGA being the TCGA, they have lots of data to manage and are constantly re-organising the data. The original files that I downloaded are now in the GDC Legacy Archive.

HERE I have configured a link for all legacy endometrial cancer clinical data in 'biotab' format as they call it. These legacy files have 3 names for each column, one being the CDE ID. You can configure for other cancers using the Cases and Files tabs (at left)

The updated GDC Data Portal no longer appears to supply these types of files and instead provides the data as XML or JSON files for each individual patient. I do know that there were inconsistency issues with using the legacy biotab files though. It is nevertheless handy just to have them in order to look up the CDE IDs.

For TCGAbiolinks, CDE IDs appear around four-fifths of the way down on this page: https://bioconductor.org/packages/devel/bioc/vignettes/TCGAbiolinks/inst/doc/clinical.html#get_legacy_clinical_data However, they don't explicitly mention CDE IDs.

Hope that this helps.

[Fdota]

Thanks a lot, they are very helpful!