Interpreting GISTIC2 output

Question

[vctrm67]

I am wondering how to get GISTIC to output, or how to post-process, data that shows copy number change per gene relative to a threshold for each sample.

According to the GISTIC documentation, the "all_lesions" file should contain actual copy number information for each sample at each amplification peak. However, I don't see specific copy number information; the closest thing I see is lines like this:

Amplification Peak  1 - CN values   1p36.33     chr1:1-658025(probes 1:58)                              chr1:1-657787(probes 1:57)                              chr1:1-909561(probes 1:62)                              1.5606e-05   1.5606e-05     Actual Copy Change Given    -0.57388    -0.6114 0.77449 -0.063946   0.12539 -0.19065    -0.08914    2.6763  0.025965    -0.62321    -0.41281    0.020825    0.42005 0.076114    0.03206 1.6868  0.67058 0.65499 -0.022245   0.40872 0.87875 2.8223  0.76756 1.2689  1.7202  3.6569  0.28766 -0.15508    -0.097146   0.7341  -0.55353    -0.80629    -0.7621 0.061591    0.029526    3.6569  0.70752 0.26434 0.20293 0.078778

However, these don't seem to be actual copy number values, and I'm not sure how to interpret them.

I checked the amplifications and deletions file, and those just tell me which genes correspond to which peak region, with no copy number information.

[Ram]

Please ask the gistic-forum - that's the help site recommended on the GISTIC website.

[vctrm67]

I would ask but they are not very responsive.

[Ram]

I'd ask just in case. Some threads get busy really quick while others go stale.

Answer 1

[MatthewP]

seg.CN data in input seg file is log2(copy number) - 1, so you need to convert data from "all_lesions" file back to raw copy number.
"actual copy number" of the first sample in your example data is "-0.57388" then you can get raw copy number by:

> seg_cn <- -0.57388
> raw_cn <- 2 ^ (seg_cn + 1)
> raw_cn
[1] 1.343615

Note: assuming ploidy is 2.