Parameters Sent To DNAcopy Functions

Dario Strbenac

unread,

Sep 29, 2010, 11:00:10 PM9/29/10

to aroma-af...@googlegroups.com

Hello,

I remember reading a while ago that you can pass in additional parameters to CbsModel, and they will get passed onto DNAcopy functions. However, it doesn't seem to be working for me. I don't want any segments less than 5 probes wide. However, the 8th segment is only 2 wide.

I don't think the warnings are related to my question, but here they are, anyway :

> warnings()
Warning messages:
1: In log(M, base = 2) : NaNs produced
2: In log(A, base = 2) : NaNs produced
3: In DNAcopy::CNA(genomdat = data$y, chrom = data$chromosome, ... :
array has repeated maploc positions

4: In log(M, base = 2) : NaNs produced
5: In log(A, base = 2) : NaNs produced
6: In DNAcopy::CNA(genomdat = data$y, chrom = data$chromosome, ... :
array has repeated maploc positions
... ... ...

I'm using aroma.affymetrix 1.7.0 on R 2.12.0 alpha.

--------------------------------------
Dario Strbenac
Research Assistant
Cancer Epigenetics
Garvan Institute of Medical Research
Darlinghurst NSW 2010
Australia

Henrik Bengtsson

unread,

Sep 29, 2010, 11:14:03 PM9/29/10

to aroma-affymetrix

Hi.

On Wed, Sep 29, 2010 at 8:00 PM, Dario Strbenac
<D.Str...@garvan.org.au> wrote:
> Hello,
>
> I remember reading a while ago that you can pass in additional parameters to CbsModel, and they will get passed onto DNAcopy functions. However, it doesn't seem to be working for me. I don't want any segments less than 5 probes wide. However, the 8th segment is only 2 wide.
>
>> model <- CbsModel(extract(normalisedCels, 1), extract(normalisedCels, 2), min.width = 5)
>> fit(model, force = TRUE)

Your expectation that you should specify the extra parameters in the
setup of the CbsModel object follows the overall style of the aroma
framework. However, in this particular case we haven't implemented
passing parameters that way. However, a workaround is to do it via
the fit() call instead. In your case, you would do:

model <- CbsModel(extract(normalisedCels, 1), extract(normalisedCels, 2));
fit(model, min.width=5, force=TRUE);

Hope this helps

Henrik

> --
> When reporting problems on aroma.affymetrix, make sure 1) to run the latest version of the package, 2) to report the output of sessionInfo() and traceback(), and 3) to post a complete code example.
>
>
> You received this message because you are subscribed to the Google Groups "aroma.affymetrix" group with website http://www.aroma-project.org/.
> To post to this group, send email to aroma-af...@googlegroups.com
> To unsubscribe and other options, go to http://www.aroma-project.org/forum/
>

Dario Strbenac

unread,

Sep 29, 2010, 11:20:04 PM9/29/10

to aroma-af...@googlegroups.com

Thanks for this.

Henrik Bengtsson

unread,

Nov 28, 2010, 5:53:13 PM11/28/10

to aroma-affymetrix

Just a follow up on this thread and close it:

Starting with aroma.core v1.8.0 (2010-11-07) it is possible to specify
additional parameters when setting up the segmentation model,
parameters that will be passed to the actual segmentation method. For
instance, (instead of having to pass them as arguments to fit(), the
preferred way is now to do:

seg <- CbsModel(ds, min.width=5);
fit(seg);

This means that also:

seg <- CbsModel(ds, min.width=5);
ce <- ChromosomeExplorer(seg);
process(seg);

will also work (where fit() is called implicitly by process()).

Note that the aroma framework only pass on these optional arguments to
the underlying segmentation method. In order to find out which they
are and what they do, see the specific segmentation method, e.g.
help("segment", package="DNAcopy").