[BioC] aCGH and Affymetrix

[Sean Davis]

On Tuesday 17 October 2006 23:06, Di Wu wrote:
> Dear All,
>
> I have some data contain both aCGH data and affymetrix data. Same samples
> aere used to generate the two types of arrays. aCGH arrays used are Agilent
> custom HD-CGH microaarays(44k). I have the package aCGH in R.
>
> I want to combine the two kinds of data together according to chromosome
> location(any other suggestion?), so that I can know the region and gene
> list which both amplified in tumor cell and other catogorical change
> between the two types of data. I can find the way to visualize the result
> along the chromosomes: but is there any way to calculate the results by
> numbers and report the region or gene list in a numerical list? I don't
> want to circle all the dot I am interested in by hand, to enlarge them, to
> see what gene are inside.
>
> Anybody with aCGH data analysis experience? please help me out.

I would do something like the following:

1)  Segment the CGH data using any of a number of CGH segmentation methods.  
The snapCGH package makes it easy to compare methods and has some plotting 
features, etc.  Doing so results in (nearly) every region of the genome 
having an estimated copy number (actually log-ratio).  

2)  Using the data from (1), determine the copy number associated with the 
genomic locations of the affymetrix probes (yes, based on genomic location).  
Then, you have paired copy number and gene expression for each affy probe.  

With segmented CGH data and gene expression, many more analyses are possible, 
but the hypotheses to be tested need to be driven by the experiment and 
experimental design.

Sean