Add copy number analysis for input

[simonelsasser]

I'm thinking it would be a nice QC to run a CNV analysis on every input file (BAM or bigwig).
Since we work with mouse ESC often, they are known to acquire some duplications. ChIP input is not ideal for calling CNV but if we focus on large regions (100kb+) I think this still works with noisy ChIP input.

[marcelm]

There are various methods to do this, but it appears to me that here only the methods that use read depth would work. Did you have that in mind?

A couple of tools I found:

• CNVnator
• CNVkit
• ClinCNV

Do you perhaps have a dataset with a known duplication that we could use for testing this?

My knowledge of this is limited, so I asked a couple of my colleagues as well.

[simonelsasser]

yes we have, in /webexport/rAWQu8Bm8w there are are two experiments (P1, P2) with three cell lines each (J1, CM_D7, KO_H9) and they have some copy number variations that I found just by loading input into IGV,
. These are very broad duplications, so just try if the tools can pick it up