[BioC] snapCGH

[Sean Davis]

On Monday 20 November 2006 10:55, João Fadista wrote:
>  Hi everyone,
>
> As I read the "snapCGH: Segmentation, Normalization and Processing of aCGH
> Data User´s Guide" I became really excited with all the features in it to
> analyse CGH data and because it is designed to be used in conjunction with
> limma package, which I have already been using. I have done the practicals
> and browsed the main functions using the data given in the package.
>
> After this stage I wanted to deal with a real data set so I downloaded a
> CGH experiment from GEO (Gene Expression Omnibus) and put it on R workspace
> using the GEOquery package. After that I converted the GEO DataSet into an
> MAList to be able to use the data with snapCGH package.
>
> Despite of this, when I used the function processCGH it gave me an error:
> > MA2 <- processCGH(MA, method.of.averaging=mean, ID="ID")
>
> Error in processCGH(MA, method.of.averaging = mean, ID = "ID") : $design
> component is null
>
> So, then I managed to to make the design column, but it gave me an error, 
but a different one:
> > MA$design <- rep(1,10)
> >
> > MA2 <- processCGH(MA, method.of.averaging=mean, ID="ID")
>
> Error in order(na.last, decreasing, ...) : argument 1 is not a vector
>
>
>
> Therefore, if MA is an object of class MAList this function should work. I
> do not see what is wrong. Isn´t the snapCGH package compatible with the
> GEO datasets?
>
> There is also another thing. In the examples folder of the package, the
> clones.info file has the columns Chromosome and Position, but in the
> dataset from GEO there is only the Entrez.GeneID identifier. Do you know of
> anyway I could convert one into another?

Hi, Joao.

All the CGH methods that are available via bioconductor require a chromosome 
and basepair position.  They cannot work without these.  There are a number 
of ways to get chromosome location, but perhaps the simplest is to use the 
biomaRt package to go from gene_id to chromosome and position.  I don't think 
that you will be able to proceed without having the chromosome locations 
included in the MA$genes data frame and, although I am not sure, I would 
guess that the error is because of not having these.  Perhaps others on the 
list will confirm this.

Sean