Question about getting transcriptomic profiles from CBioPortal

[Altwegg, Kristin (NIH/NCATS) [F]]

Dear Friends at cBioPortal,

 

I am a postdoc researcher with a dual fellowship between NIH and FDA. I am working on a drug repurposing project in head and neck squamous cell carcinomas. I was on cBioPortal looking for patient data for pre- and post-treatment RNA-seq and mutations (large screen grab). I did find the data on the site (smaller screen grab) but when I tried to download it, I only got a small file with the drug names. For example, on the smaller screen grab for Treatment per Sample (pre/post) FLUOROURACIL – Pre has 4,931. What does this number mean? Is there tumor sequencing data? Can I get access to this information? How can I obtain the profiles of the tumors pre- and post-treatment if it is available?

 

Thank you in advance for your time and assistance.

 

Kristin Altwegg

 

 

Kristin Ann Altwegg, Ph.D

TSIF Postdoctoral Fellow

Early Translation Branch | Division of Preclinical Innovation

National Center for Advancing Translational Sciences (NCATS)

National Institutes of Health, HHS

Lab B1, Office 1034D

9800 Medical Center Drive

Rockville, MD 20850

Email: kristin...@nih.gov

NCATS Website: http://ncats.nih.gov

 

[Tali Mazor]

Hi Kristin,

There are a few things happening here, including that you've identified a bug in our website. Please bear with me as I try to work through this.

First, I see in your screenshot that you are looking at a combined study. Is this the set of studies you combined?

If so, I do need to draw your attention to two things: (1) these are all TCGA studies, which means the same samples exist in each study. I suggest you select just one of these studies for your analysis. (2) These studies are a mix of genome builds (hg19 and hg38) which cBioPortal does not yet fully support, so again I would advise picking just one of these studies.

It then looks like you selected based on "Cancer Type". In doing so, you filtered down to just one study - you can see this by looking at the "Cancer Studies" pie chart, which indicates all the samples remaining after filtering are from the GDC cohort.

If you look at the GDC cohort on its own, you'll see that there is no treatment data available: https://www.cbioportal.org/study/summary?id=hnsc_tcga_gdc  It seems there is a bug that is leading to inaccurate numbers in the "Treatment per Patient" and "Treatment per Sampel" charts in the combined study view. I have reported this and you can track our progress on resolving the issue here: https://github.com/cBioPortal/cbioportal/issues/11532

The PanCancer Atlas is the only of these cohorts with treatment data. I suggest running your analysis in just this cohort: https://www.cbioportal.org/study/summary?id=hnsc_tcga_pan_can_atlas_2018

As for the numbers in the "Treatment per Sample" chart, those indicate the number of samples  in the cohort that were acquired before or after the patient received a particular treatment. You'll notice that almost all samples were acquired prior to treatment ("pre"), and very few samples are from after a treatment ("post"). This is because TCGA focused on studying tumors at diagnosis, and therefore there are very few samples acquired after treatments were given.

There are a few ways you can access data about samples pre/post treatment, depending on exactly what you are trying to do. You can filter in study view to eg pre-cisplatin samples and explore those. If you want to capture the sample IDs associated with those samples, you can use the download button next to the 'Custom Selection' button to download the clinical data, including sample IDs, of the currently filtered samples. Another option is to click through to patient view (click the person icon next to the number of samples in the top right) and browse through the individual cases.

Another approach is you can click the check boxes in the Treatment per Sample table, but rather than clicking "Select Samples" at the bottom, you can instead hover over the top right corner of the chart and select "Compare Groups" in the menu that appears - this will allow you to compare mRNA expression between groups, among other things.

Or, if you simply want the full genomic/expression profiles of all samples, you can download that by clicking the download button at the top of the page next to the study name.

I hope this helps. Feel free to reach out if you have additional questions.

-Tali

--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cbioportal/SA1PR09MB750412F9E163F38E3B0970DDF088A%40SA1PR09MB7504.namprd09.prod.outlook.com.

[Altwegg, Kristin (NIH/NCATS) [F]]

Dear Tali,

 

This is great advice! Thank you so much for the detailed email, I am incredibly appreciative. I will take your suggestion and start an analysis of the data you have suggested. 

 

Also, thanks for allowing me to track the progress on resolving the bug issue. I am a few weeks into learning to code in R and this is all very interesting (and exciting) to me.

 

Warmest regards,

 

Kristin

 

From: Tali Mazor <tma...@ds.dfci.harvard.edu>
Sent: Thursday, May 8, 2025 10:28 AM
To: Altwegg, Kristin (NIH/NCATS) [F] <kristin...@nih.gov>
Cc: cbiop...@googlegroups.com
Subject: [EXTERNAL] Re: [cbioportal] Question about getting transcriptomic profiles from CBioPortal

 

Hi Kristin,

 

There are a few things happening here, including that you've identified a bug in our website. Please bear with me as I try to work through this.

 

First, I see in your screenshot that you are looking at a combined study. Is this the set of studies you combined?

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and are confident the content is safe.