[BioC] SNP analysis

[Holger Schwender]

Hi JB,

apart from many general classification methods that can also be applied to SNPs (packages for these procedures are available on CRAN, i.e. http://cran.r-project.org -- one example is the package randomForest containing functions for the classification method Random Forests) there is a package called logicFS available at BioConductor that uses logic regression for classification (logic regression has been developed in the context of SNP data) and also provides a function called mlogreg for multi-class problems. Logic regression itself can of course also be used for classification (and regression) and is available in the package LogicReg on CRAN (if you install logicFS, LogicReg will automatically be installed).

There is also a package on CRAN called scrime which contains several tools for analyzing SNP data. Besides other things it contains classification approaches for categorical data based on nearest neighbors and shrunken nearest neighbors (pamCat).

If you are just interested in identifying SNPs that differ between classes you can also use the functions in scrime for performing rowwise multiple testing with Pearson's ChiSquare test or the Cochran-Armitage trend test. The former can also be used in the current release version of the package siggenes available on BioConductor. The development version additionally contains the possibility of employing the latter test.

Best,
Holger

PS: This is a pretty biased selection of packages, as you will notice when taking a look on the authors of the mentioned packages. For other R packages concerned with genetics, see http://cran.r-project.org/web/views/Genetics.html

-------- Original-Nachricht --------
> Datum: Thu, 12 Feb 2009 15:39:28 +0100
> Von: Jean-Baptiste Poullet <jean-baptistepoullet at dnavision.be>
> An: BioC <bioconductor at stat.math.ethz.ch>
> Betreff: [BioC] SNP analysis

> Hi, 
> 
> I'm very new in Bioconductor and more largely in genetics. I was
> wondering whether there was something usable in Bioconductor for
> analyzing breed specific SNPs (in particular for classifying individuals
> among populations). 
> 
> Thanks in advance. 
> 
> JB
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives:
> http://news.gmane.org/gmane.science.biology.informatics.conductor

--