[BioC] segmentation for Affy SNP arrays

[Rob Scharpf]

> Message: 1
> Date: Wed, 19 Sep 2007 15:26:18 +0200
> From: "Benjamin Otto" <b.otto at uke.uni-hamburg.de>
> Subject: Re: [BioC] segmentation for Affy SNP arrays
> To: "'Ramon Diaz-Uriarte'" <rdiaz at cnio.es>,
> 	<bioconductor at stat.math.ethz.ch>
> Message-ID: <001501c7fac0$b0e0b610$9f05a20a at matrix.com>
> Content-Type: text/plain;  charset="utf-8"
>
> Hi Yu,
>
> You might also have a look at Dchip, CNAG and GEMCA. Theye are no R  
> packages
> but should be able to calculate the CNs for you up to the 500k chips.

Once you have SNP-level summaries of genotype and/or copy number, the  
R package VanillaICE can be used to fit a hidden Markov model.  The  
underlying states (e.g., deletion, amplification, etc) are determined  
by whether genotype, copy number, or both are available (the  
available SNP-level summaries define the class of the object).  We've  
only tested on the Affymetrix 100k, 500k and Illumina platforms.  But  
as long as you have the physical position of a SNP and a  
corresponding estimate of copy number and/or genotype, the 'Vanilla'  
HMM should be platform-independent.

Rob

> -----Urspr?ngliche Nachricht-----
> Von: bioconductor-bounces at stat.math.ethz.ch
> [mailto:bioconductor-bounces at stat.math.ethz.ch] Im Auftrag von Ramon
> Diaz-Uriarte
> Gesendet: Wednesday, September 19, 2007 11:10 AM
> An: bioconductor at stat.math.ethz.ch
> Betreff: Re: [BioC] segmentation for Affy SNP arrays
>
> Dear Yu,
>
>
> The packages that you mention (and some others, too) can be used  
> with Affy
> SNP arrays. The paper by Yu et al., 2007, BMC Bioinformatics,8: 145+
> (http://www.biomedcentral.com/1471-2105/8/145/abstract), for example,
> compares their performance. (Please note that we have some problems  
> with the
> way our RJaCGH package was used and we are working on a reply. I  
> can send
> further details off-list).
>
> Best,
>
> R.
>