Copy Number Amplification

[PJ]

Hello,

I have a few questions regarding the copy number analysis.

1) What does 1 & 2 mean in terms of number of copies. I know 2 is high
level amplification and 1 is low level or moderate gains but how do
you draw the cutoff.

2) Do you plot survival curves based on 1 & 2 vs 0 (diploid) or just
2, etc...

3) We may partition the data differently based on different CNA,
mutations, and create different groups for KM plots...I'm assuming
you're using survdiff with rho=0, but just want to know to be
consistent

4) Which brings me to the 4th question, if we're using calls from your
portal as to amplified and diploid etc... for publication purposes I
know to cite TCGA but would also like to cite this group as well, can
you provide me with the appropriate citation. I'm sure its somewhere
on this portal but since I'm asking a bunch of other questions thought
I would throw this in there to be sure.

Thanks a bunch in advance. This is a great portal and the
visualizations are impressive.

Cheers,

Pichai

[Nikolaus Schultz]

Dear Pichai,

Thank you for your interest in the cBio Portal.

Below are the answers to your questions:

1)

Copy-number calls are defined by GISTIC or RAE, depending on the cancer study, and it is not straightforward to directly convert 1 and 2 into number of copies. A gain (1) typically means a lower-level arm-level gain, so a few extra copies. Amplification (2) refers to higher level, usually focal amplifications (many extra copies, ~5 and up).

2)

Survival curves are plotted based on default or user-defined alteration levels. The default copy-number alterations are homozygous deletions (-2) and high level amplifications (2). But you can include low-level gains (1) in the analysis by using the OncoQuery Language. For example, enter CDK4: GAIN AMP in the query field and samples with either a gain or an amplification will be considered as altered.

3)

Yes, we are using survdiff with rho=0.

4)

In your publications, please cite our paper that just came out:

Cancer Discov. 2012 May;2(5):401-4.

http://www.ncbi.nlm.nih.gov/pubmed/22588877

Please let us know if you have any other questions or suggestions.

Niki.