CNA differences when querying with or w/o HOMDEL HETLOSS

[Friederike Hanssen]

Hi!

I have been looking at a few genes, and am a bit puzzled on the reported numbers for CNA:

My steps are the following:

1.  Tab ‘Query’

2. Select ‘Liver’

3. Select ‘Liver Hepatocellular Carcinoma (TCGA, PanCancer Atlas’

4. Press ‘Query by Gene’

5. In the field ‘Enter genes’, we tried 2 approaches:

a) KDM6A

        b) KDM6A: HOMDEL HETLOSS

With approach a) the results look like this:

 

With approach b), like this:

We are puzzled by the fact that in the first image only deep deletions are shown, but no shallow deletions. We would have expected, all alterations to be displayed. 

We are unsure now about what is ’true’. The number of patients remains the same with 353. Do we need to specify all CNA types, if we want them to be displayed? 

Additionally, we have found for quite a few patients, when we click on them, that no CNA or the gene we are investigating is not displayed in their table at the bottom. 

For example the patient TCGA-BC-A5W4 is labeled as ’Shallow Deletion’ in the b) plot. However, when searching through the ‘Copy Number Alteration’ Table at the bottom, the gene is not listed in it. Others are greyed out in the a) or b) plot - labeled as ’no alterations’ - but when clicking on them nothing seems to be annotated on the X/Y chromsome. Are these truly not alterated or are they not analysed? And what does shallow deletion for a gene on the X chromosome mean for male patients? Since they only have a single X chromosome, we were wondering whether this is always labeled as Shallow deletion? However, this patient, TCGA-2Y-A9GW, is classified as Deep Deletion, even though he is male. 

Last, but not least we tried the following steps:

1. Select ‘Data Sets’ Tab
2. Search ‘Liver’
3. Select ‘Liver Hepatocellular Carcinoma (TCGA, PanCancer Atlas)’
4. Select only HCC patients (upper left box in view)
5. Go to ‘CN Segments’
6. Query ‘KDM6A’ 
   
   
   

We would appreciate some feedback and/or help in how to get the correct plot here. 

Please let me know, if you need any further information from me to clarify my question. 

Best regards,

Friederike Hanssen

--
M.Sc. Friederike Hanssen

PhD Student
Quantitative Biology Center (QBiC)
Eberhard Karls Universität Tübingen
Auf der Morgenstelle 10, D-72076 Tübingen

email: friederik...@qbic.uni-tuebingen.de
http://qbic.life

[Sjoerd van Hagen]

Hi Friederike,

In your first example, only deep deletions are shown by default indeed. This is the expected behaviour and as you found you can indeed also get the shallow deletions. The reason for taking the deep deletions as the default is that with a shallow deletion, quite often the remaining copy is sufficient. Of course there are exceptions to this as with almost anything in life sciences.

In your second example, when I follow the first 5 steps, I end up with this:

Now, if you want to see the CNA segments around KDM6A, you need to type that gene in the box inside the IGV, which now says 'all'. If you put the gene in the box at the top right and hit query, you will actually go to a different view, the one in your screenshot, which is not what you want (I think).

I hope this helps.

Best,

Sjoerd.

---

Sjoerd van Hagen

Team Lead cBioPortal & Open Targets

E sjo...@thehyve.nl

T +31 30 700 9713

W thehyve.nl

    

--
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/C619C06E-371E-4E52-95BB-C81B673CF4A5%40qbic.uni-tuebingen.de.

[Sjoerd van Hagen]

Hi Friederike,

You are welcome!

Yes, no alteration means no alteration as defined by what was provided in the query box. Patients may still have alterations in other genes, or of other types, and for continuous variables like expression z-scores you can also set different thresholds.

Not analyzed is defined by gene panels but for the TCGA studies these were not used so you can assume that all genes were analyzed and not colored means that there is no alteration according to the definition in the query box.

I am not sure how GISTIC is calling these CNA but there tends to be a lot of debate about it. I suppose that for a shallow deletion there must still be a significant number of reads for that gene in the sample but fewer than the number of reads you would expect for unaltered genes. This would suggest that in part of the tumor there may be a deep deletion while in other parts the gene is still present. With a heterogeneous tumor you will often see results that appear strange at first but can be explained by that bulk sequencing gives you an average over the tumor. For instance, you can also find mutations in genes that are called as deep deletion. Perhaps Niki knows more.

The CNA view in the patient view is filtered based on known driver genes, and I guess also on deep deletions. The reason for this is that often long stretches of CNA are deleted and then you would have hundreds or thousands of genes that are not relevant. You can see this nicely illustrated if you do the comparison between patients with a KDM6A deletion versus unaltered. You will see a long list of genes that are highly enriched in the altered group because they have deletions of adjacent genes.  You can see this in the cytoband column: they are all on the short arm of the X-chromosome.

I hope this helps,

Sjoerd.

---

Sjoerd van Hagen

Team Lead cBioPortal & Open Targets

E sjo...@thehyve.nl

T +31 30 700 9713

W thehyve.nl

    

On Mon, Mar 15, 2021 at 10:53 PM Friederike Hanssen <friederik...@qbic.uni-tuebingen.de> wrote:

Hi Sjoerd,

Thank you for your prompt reply.

To clarify: the patients with `no alteration` then really are just not annotated/colored in a certain view, but may as well have some other type of alteration, right?

So really, we have 28% of patients with some type of deletion (as shown in plot b), even if not shown in my plot a)).  Is there a good way to distinguish between patients with truly ’no alteration’ vs ’not colored’ vs ’not analyzed'?

Additionally, what does `shallow` deletion mean for male patients for a gene on the X chromosome? Is Deep Deletion = HOMDEL and Shallow Deletion = HETLOSS? How come male patients are then present in both groups?

And lastly, why are for many patients no CNA listed in the table at the bottom of the patient’s page? 

Thank you for your help. I really appreciate it.

Best,

Friederike

 

--
M.Sc. Friederike Hanssen

PhD Student
Quantitative Biology Center (QBiC)
Eberhard Karls Universität Tübingen
Auf der Morgenstelle 10, D-72076 Tübingen

email: friederik...@qbic.uni-tuebingen.de
http://qbic.life

On 15. Mar 2021, at 22:32, Sjoerd van Hagen <sjo...@thehyve.nl> wrote:

Hi Friederike,

In your first example, only deep deletions are shown by default indeed. This is the expected behaviour and as you found you can indeed also get the shallow deletions. The reason for taking the deep deletions as the default is that with a shallow deletion, quite often the remaining copy is sufficient. Of course there are exceptions to this as with almost anything in life sciences.

In your second example, when I follow the first 5 steps, I end up with this:

 <Screenshot 2021-03-15 at 10.30.41.png>

With approach b), like this:

<Screenshot 2021-03-15 at 10.31.15.png>

We are puzzled by the fact that in the first image only deep deletions are shown, but no shallow deletions. We would have expected, all alterations to be displayed. 

We are unsure now about what is ’true’. The number of patients remains the same with 353. Do we need to specify all CNA types, if we want them to be displayed? 

Additionally, we have found for quite a few patients, when we click on them, that no CNA or the gene we are investigating is not displayed in their table at the bottom. 

For example the patient TCGA-BC-A5W4 is labeled as ’Shallow Deletion’ in the b) plot. However, when searching through the ‘Copy Number Alteration’ Table at the bottom, the gene is not listed in it. Others are greyed out in the a) or b) plot - labeled as ’no alterations’ - but when clicking on them nothing seems to be annotated on the X/Y chromsome. Are these truly not alterated or are they not analysed? And what does shallow deletion for a gene on the X chromosome mean for male patients? Since they only have a single X chromosome, we were wondering whether this is always labeled as Shallow deletion? However, this patient, TCGA-2Y-A9GW, is classified as Deep Deletion, even though he is male. 

Last, but not least we tried the following steps:

1. Select ‘Data Sets’ Tab
2. Search ‘Liver’
3. Select ‘Liver Hepatocellular Carcinoma (TCGA, PanCancer Atlas)’
4. Select only HCC patients (upper left box in view)
5. Go to ‘CN Segments’

1. Query ‘KDM6A’ 
   <Screenshot 2021-03-15 at 10.44.31.png>
   
   

We would appreciate some feedback and/or help in how to get the correct plot here. 

Please let me know, if you need any further information from me to clarify my question. 

Best regards,

Friederike Hanssen

--
M.Sc. Friederike Hanssen

PhD Student
Quantitative Biology Center (QBiC)
Eberhard Karls Universität Tübingen
Auf der Morgenstelle 10, D-72076 Tübingen

email: friederik...@qbic.uni-tuebingen.de
http://qbic.life

[Nikolaus Schultz]

Hi Friederike,

These gene-specific copy-number calls have always been very tricky, and they are almost impossible to get right 100% of the time using fully automated methods. 

Just to make sure we are all on the same page, here are a couple of basic facts about the CNAs in TCGA samples in cBioPortal:

 - copy-number in TCGA is derived from Affymetrix SNP 6 arrays using the GISTIC 2 algorithm

 - by default, cBioPortal only shows high level amplifications and deep deletions in OncoPrints. You can use OQL to show the other events. However, the patient view only shows deep deletions and amplifications

 - for calling shallow deletions, GISTIC uses the same threshold across all samples

 - for calling deep deletions, GISTIC uses a sample-specific threshold in an attempt to correct for purity

 - GISTIC does not correct for ploidy, meaning that whole genome duplicated samples are very tricky to get right (one of the main reasons we use shallow vs deep as opposed to heterozygous vs homozygous)

 - sub clonal deep deletions would probably be called shallow deletions

 - partial deep deletions of a gene are also called deep deletions - in these cases, you can observe a mutation even in a gene that is called deleted

 - it is also possible that a gene is incorrectly called “deep deletion”, and that’s why you observe a mutation in that same sample

 - I don’t think GISTIC treats the X chromosome any different, but I agree with you that different rules should be used here, especially in men 

So to answer your specific question, there are a couple of different explanations for heterozygous losses in X chromosome genes in men:

 - the deletion is sub-clonal

 - the tumor is whole-genome duplicated, and only one of two copies are lost

 - the algorithm made an incorrect call

Does this make sense?

I would suggest to inspect the segmented copy-number data more closely in IGV to get a better sense of what happens to your gene of interest in your specific cancer type.

I hope this helps.

Niki.

On Apr 15, 2021, at 5:52 AM, Friederike Hanssen <friederik...@qbic.uni-tuebingen.de> wrote:

Hi Sjoerd and Nikolaus,

I wanted to follow up on the question of deep deletions vs shallow deletions in male patients on the sex chromosomes. You already mentioned the possibility that in parts of the tumor  complete deletion took place.

I am still not entirely sure what it then means that male patients for a gene on the X chromosome can show shallow AND deep deletions, especially since you mentioned that even in patients with deep deletions, mutations can be found. Could you possibly elaborate? 

Best regards,