Several questions

[Rotraud Wieser]

To whom it may concern:

I have a few questions re cbioportal:

1) Under "Genomic alterations", to which basis are the log-ratios? log2?

2) Under the "Annotate data/Filter data button": for the definition of
drivers vs. VUS, the pre-selected choice is OncoKB and Hotspots. Are
this linked via a logical AND or an OR?.

And what does "COSMIC >=0" mean?

3) Is there any list of abbreviations; e.g. CNA, VUS, or,
study-specific: TMB?

I happen to know what these mean, but also came across some others whose
meaning was entirely unclear to me.

Any quick way to find out what these abbreviations mean?

Thanks in advance, best, Rotraud


--
Ao. Univ. Prof. Mag. Dr. Rotraud Wieser
Division of Oncology
Department of Medicine I
Medical University of Vienna
Waehringer Guertel 18-20, 1090 Vienna, Austria

Tel: +43 (0)1 40400 – 73779
Email: rotraud...@meduniwien.ac.at
https://innere-med-1.meduniwien.ac.at/onkologie/

Legal notice: This message and any attachment(s) may contain confidential/privileged information. If you are not the intended recipient, please email or call the sender
and delete this message and any attachment(s) from your system. Any unauthorised copying,
disclosure or distribution of the material in this email is strictly forbidden.

Rechtlicher Hinweis: Diese E-Mail samt Anhängen könnte vertrauliche/rechtlich geschützte
Informationen enthalten. Wenn Sie nicht der richtige Empfänger sind, informieren Sie bitte
sofort den Absender und löschen diese E-Mail samt Anhängen. Das unerlaubte Kopieren oder
Weitergeben dieser E Mail ist nicht gestattet.

[JJ Gao]

Hi Rotraud,

Thanks for contacting us.

1)  It's log2 in the Genomic Alterations tab of the Comparison page/tab

2) It's OR logic, ie. a variant will be annotated as "putative driver" if it is oncogenic/likely oncogenic based on OncoKB OR it is a cancer hotspot. "COSMIC > N" means "occurs in more than N samples in COSMIC" (note: we use a out of date COSMIC release due to data license)

3) Unfortunately, we don't. I have created a ticket for that: https://github.com/cBioPortal/cbioportal/issues/9223. Please let us know if you come across any abbreviations that you don't understand.

Best,

-JJ

--
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 on the web visit https://groups.google.com/d/msgid/cbioportal/9fe994b7-6bb4-1f92-cc20-30da602ae712%40meduniwien.ac.at.

[Rotraud Wieser]

Hi JJ,

thanks for the quick and informative reply.

I really love cbioportal, and the more I learn about it, the more I am impressed with it.

Nevertheless, here is a few suggestions for improvement:

1) name the x-axis log2 ratio instead of log ratio

2) add an OR between the different options for the definition of VUS

3) in an ideal world, I would hover over an abbreviation and see its definition. In an even "more ideal" world, I could hover over any parameter and see a brief description of it. I realize, though, that it would be a lot of work to implement this.

In any case, I will try to collect more abbreviations as I go along and send you a list.

THanks again, best, Rotraud

[JJ Gao]

Thanks for the feedback!

1) Will do: https://github.com/cBioPortal/cbioportal/issues/9227

2) We'll need to think a bit more about this. Please let us know if you have a specific use case.

3) Agreed. We'll pay more attention when designing the UI on that. For clinical attributes, we have descriptions, but it is probably not displayed everywhere it's used.

-JJ

[Rotraud Wieser]

Hi JJ,

thanks, I highly appreciate it.

Re point 2: maybe there is some misunderstanding.  All I suggest is to make clear on the website that the logical connector is an "OR", in whichever way you think this would be appropriate.

However, I fully realize that the cbioportal team has thought a lot more about cbioportal than I have, and if for some reason that does not appear feasible, never mind.

Rotraud

[JJ Gao]

Got it - I have create another ticket for that: https://github.com/cBioPortal/cbioportal/issues/9228

Thank you very much again for your input - they are important for us to improve! Please let us know if you have any more questions when using cBioPortal.

-JJ

[JJ Gao]

Hi Rotraud,

Sorry for the late reply. Please cc cbiop...@googlegroups.com for prompt response.

1. Mutation_Count is the number of mutations detected in the sample. The latter two TMB are per Mb. We usually count only nonsynonymous mutations but there might be some specific data in certain studies. If you find inconsistency, please send us the study IDs for us to check.

2. NA means the sample was not profiled (data unavailable).

3. The sample matrix from download tab? If so, yes, it's a summary of all alterations queried - really a different representation of oncoprint. 

4a. It is possible that multiple mutations occur to the same allele of a gene; and also -1 CNA does not necessarily mean there is only one copy - e.g. a whole-genome duplicated sample may have two copies and still be counted as copy number lost. 

4b. I saw some KRAS mutant case with KRAS amplification in the cohort (https://bit.ly/3FODxho), but it's only a small fraction, not all...

Best,

-JJ

On Tue, Jan 18, 2022 at 9:51 AM Rotraud Wieser <rotraud...@meduniwien.ac.at> wrote:

Dear JJ,

still studying cbioportal (for teaching), I came across a few more questions:

1) I assume, Mutation_Count, Tumor_Mutation_Burden, and TMB_nonsynonymous are expressed per Mb of DNA?

What is the difference between the first two of these?

2) In downloads of CNA and SV data, what does "NA" mean? Not applicable, not analyzed, not present....?

3) What exactly is the downloadable "sample matrix"? Does it summarize mutations, CNAs, and SV data?

4) When I tried to figure out Q3 myself, I came across the following: In the PanLung study (1144 patients), the patient in line 27 of the downloaded file has 2 different p53 mutations (1 stop, 1 missense downstream of the stop), plus a -1 in the CNA table. How can there be e compound heterozygous mutation in a patient with a copy number loss? S/he should have only 1 p53 allele.... Also, it seemed like all patients that had a mutation in any gene (incl. oncogenic RAS mutations) also had a CNV in that same gene.

Can you help again?

Thanks in advance, Rotraud

[Rotraud Wieser]

Hi JJ,

thanks again.

Best, Rotraud