[BioC] snapCGH; export results of summarize.clones() function from aCGH

[Ian Roberts]

Dear Steven

> I looked at older versions of snapCGH, and
> indeed version 1.1.0 had a function
> summarizeClones()
> which has indeed disappeared :(

It is a piity - but I know there is a reluctance to implement these functions 
from a Stats perspective.  E.g SeLists define large regions of interests.  The 
worry is that exporting the clones within such a region somehow validates the 
segmented feature as being the final result.  From a Biologists view, we need 
the clone names because they are a short cut to getting back into the genome 
and enabling us to validate any imbalances by FISH or qPCR.  They also make 
comparisons across samples easier for the non-stats minded people (me).

> I do not see a straightforward way of
> converting a SegList object to an aCGH type
> object, I'm not up to date on the definitions
> of the various components of snapCGH's SegList 
> and aCGH objects.  For example, summarize.clones()
> seems to work with kb units for clone positions
> but snapCGH works with Mb units so some care
> is needed to convert objects.

It took me 12 hr yesterday to export and transform the MA$M and MA$genes data 
frames from snapCGH and get them in to aCGH compliant objects.  I've made notes 
and will write a recipe when I am happy that what I've done is ok.  Needless to 
say, it took me 11 hr to spot the extra white space introduced during the 
conversion ...  One problem at the moment - snapCGH handles dye-swap easily 
because of LiMMAs modelMatrix function --> any idea how to do the same with aCGH

> 
> The function nudSegmentation() in snapCGH appears to
> generate the type of data from which output similar 
> to that from summarize.clones() could be obtained
> (see the "regions" component of nudSegmentation
> output). 

I will see whether this offers a short cut.

Thanks for yout input.  I feel I am making progress.
best wishes,
Ian