[BioC] mixed models analysis with limma

[Gordon K Smyth]

You have to specify the design matrix to duplicateCorrelation().

Gordon

On Mon, 28 Sep 2009, Christian Brière wrote:

> Dear Gordon,
> Thank you for your help. This exactly what I need I think.
> May I ask you some more help ? I would like to know whether the script below 
> is correct :
>
> There are 4 biological experiments : F, H, O and CB
> and there is 1 treatment with three levels: T , C and CL
>
> I'd like to know genes that are differentially expressed between "C" and "T", 
> and between "CL" and "C":
>
>> #
>> # Normalized data are in Dnorm
>> exp<-c("F","F","F","H","H","H","H","O","O","O","CB","CB","CB","CB")
>> corfit<-duplicateCorrelation(Dnorm, block=exp)
>> corfit$consensus
> [1] -0.07705255
>> tr<-c("T","C","CL","T","C","CL","T","C","CL","T","C","CL")
>> tr<-factor(tr, levels=c("T","C","CL"))
>> design<-model.matrix(~tr)
>> fit<-lmFit(Dnorm,design, block=exp, cor=corfit$consensus)
>> fit<-eBayes(fit)
>> # comparison between treatment C and T
>> ttC<-topTable(fit, coef="trC", p.value=0.01, n=Inf)
>> dim(ttC2)
> [1] 8834    7
>> # comparison between treatment CL and C
>> cont.matrix<-cbind("CLvsC"=c(0,1,-1))
>> fitCCL<-contrasts.fit(fit2, cont.matrix)
>> fitCCL<-eBayes(fitCCL)
>> ttCCL<-topTable(fitCCL, coef=1, p.value=0.01, n="Inf")
>> dim(ttCCL)
> [1] 8798    7
>>
>
> Thank you for your help
> Christian
>
> Gordon K Smyth a écrit :
>> Dear Christian,
>> 
>> See Section 8.2 of the limma User's Guide.  The 'block' argument of 
>> duplicateCorrelation() and lmFit() handles random effects.
>> 
>> Best wishes
>> Gordon
>> 
>>> Date: Fri, 25 Sep 2009 08:33:43 +0200
>>> From: Christian Bri?re <briere at scsv.ups-tlse.fr>
>>> Subject: [BioC] mixed models analysis with limma
>>> To: bioconductor at stat.math.ethz.ch
>>> Content-Type: text/plain
>>> 
>>> Hello,
>>> 
>>> I would like to know whether it is possible to analyze microarray data
>>> from a mixed model experiment with Limma package.
>>> I have 4 independent experiments (random effect) and 3 treatments (fixed
>>> effect), and 1 microarray (monocolor) for each combination of the two
>>> factors.
>>> 
>>> Thank you for your help
>>> -- 
>>> 
>>> Christian Bri?re
>>> UMR CNRS-UPS 5546
>>> BP42617 Auzeville
>>> F-31326 Castanet-Tolosan (France)
>>> tel: +33(0)5 62 19 35 90
>>> Fax: +33(0)5 62 19 35 02
>>> E-mail: briere at scsv.ups-tlse.fr <mailto:briere at scsv.ups-tlse.fr>
>>> 
>>> http://www.scsv.ups-tlse.fr
>>> http://www.gdr2688.ups-tlse.fr <http://www.gdr2688.ups-tlse.fr/index.php>
>>> http://www.ifr40.cnrs.fr
>
>
> -- 
>
> Christian Brière
> UMR CNRS-UPS 5546
> BP42617 Auzeville
> F-31326 Castanet-Tolosan (France)
> tel: +33(0)5 62 19 35 90
> Fax: +33(0)5 62 19 35 02
> E-mail: briere at scsv.ups-tlse.fr <mailto:briere at scsv.ups-tlse.fr>
>
> http://www.scsv.ups-tlse.fr
> http://www.gdr2688.ups-tlse.fr <http://www.gdr2688.ups-tlse.fr/index.php>
> http://www.ifr40.cnrs.fr
>
>
>
>