[BioC] Question on Limma technical replicate

Gordon K Smyth smyth at wehi.EDU.AU
Tue Jul 6 04:25:08 CEST 2010


Dear Eleonora,

You need to explain what you want to test for before we can advise you on 
the correct contrast.

Also, if your dye-swaps are technical replicates, it appears that your 
experiment doesn't have any biological replication.  If you've repeated 
the RNA extraction, labelling etc for the dye-swaps then all is well.  If 
not, then the experiment is difficult to analyse from the point of view of 
statistical significance.

Best wishes
Gordon

-----------------------------------------------
Associate Professor Gordon K Smyth,
NHMRC Senior Research Fellow,
Bioinformatics Division, 
Walter and Eliza Hall Institute of Medical Research, 
1G Royal Parade, Parkville, Vic 3052, Australia.
smyth at wehi.edu.au
http://www.wehi.edu.au
http://www.statsci.org/smyth


> Date: Sun, 4 Jul 2010 15:42:13 +0200 (CEST)
> From: "Leleacquario at libero.it" <t>
> To: bioconductor at stat.math.ethz.ch
> Subject: [BioC] Question on Limma technical replicate
>
>
>
> Hi everyone,
> I am trying to analyze an   Agilent 4x44  experiment with the following
> design:
>
>> FileName  Cy3     Cy5       time   cell_line
>  1                M48R    M48C     48       MCF7
>  2                M48C    M48R     48       MCF7
>  3                M6R      M6C        6         MCF7
>  4                M6C      M6R        6         MCF7
>  5                D48R    D48C     48       MDA
>  6                D48C    D48R     48       MDA
>  7                D6R       D6C       6         MDA
>  8                D6C       D6R       6         MDA
>
>
> where: M= MCF7 cell line  while D=MDA cell line.
> The number indicates the time and the final letter (C or D) indicates the
> treatment (R) or control (C).
> After  normalization ecc ecc, I create my design file:
>
>
> MCF748 <- grep("M48",targets$Cy3)
> MCF76 <- grep("M6",targets$Cy3)
> MDA48 <- grep("D48",targets$Cy3)
> MDA6 <- grep("D6",targets$Cy3)
> 	designMCF748 <- modelMatrix(targets[MCF748,],ref="M48C")
> 	designMCF76 <- modelMatrix(targets[MCF76,],ref="M6C")
> 	designMDA48 <- modelMatrix(targets[MDA48,],ref="D48C")
> 	designMDA6 <- modelMatrix(targets[MDA6,],ref="D6C")
> 	   design <- blockDiag(designMCF748,designMCF76,designMDA48,designMDA6)
>
>
> and then:
>
> fit<- lmFit(MA, design, weights=NULL)
>    contrast.matrix <-makeContrasts(M48R,M6R,D48R,D6R, levels=design)
>    fit2<-contrasts.fit(fit,contrast.matrix)
>    efit<-eBayes(fit2)
>
>
> Now..my problem is: how to consider the dye-swap arrays?
> When I makeContrasts as above, it automatically excludes the swapped arrays.
> Even trying with the function: "duplicateCorrelation", the output is  "NA".
> for all the values. I hope someone can help me!
>
> Thanks in advance for any help.
> Eleonora

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}



More information about the Bioconductor mailing list