fitting a copy number profile to know CNV signatures

Question

[Bogdan]

Dear all,

I would like to ask for a suggestion on whether I could fit a CNA profile of a sample that I do have to known CNV signatures.

I have used the package sigminer, that seems to focus entirely on CNV signature discovery nowadays, and not on fitting of a sample CNV to known CNV profiles (that are listed on https://cancer.sanger.ac.uk/signatures/cn/)

Given the data in the format :

  *Chromosome Start.bp   End.bp modal_cn minor_cn              sample

1       chr1  1439593  2042815        3        1 EA5040545.EA5040598

2       chr1  2425333  5924715        1        0 EA5040545.EA5040598*

which package shall I use in the order to find the proxy of my sample to known CNA signatures ?

Thanks a lot,

Bogdan

Answer 1

[LauferVA]

Hi Bogdan,

Another cool question I see :-)

One option that comes to mind is to use the available code and data from the Alexandrov lab, e.g. sigprofiler, which was I think developed for this manuscript. I may not understand your goal precisely, but wouldn't either SigProfilerSingleSample or SigProfilerAssignment work, depending on exactly what you mean?

If I've slightly misunderstood, let me know, can go from there. If Im off, it seems like it might be feasible to write a quick script or two, but of course i dont know your data type, amount, etc.

Best, VL

[Bogdan]

HI Vincent,

thank you for your suggestions.I came across SigMIner https://shixiangwang.github.io/sigminer-book/analysis-supps.html ; documentation can be also found on the page : https://github.com/ShixiangWang/sigminer

I am testing it now to check whether it fits our needs.

Thanks so much,

~ Bogdan

[LauferVA]

copy that - good luck