5.4 years ago by
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
> Date: Sat, 4 Jan 2014 19:58:32 +0000
> From: Joseph Shaw <josph.sh at="" gmail.com="">
> To: Ryan <rct at="" thompsonclan.org="">
> Cc: bioconductor at r-project.org
> Subject: Re: [BioC] Design/Contrast for Two-Channel Experimental
> Hi Ryan,
> Thanks for your reply!
> It was my belief that the experimental setup would imply that the
> effect would be confounded with the biological effect - thanks for
> clarifying that this is indeed the case. However, I'm still slightly
> confused about the dye effect term; specifically, shouldn't the
> normalisation (performed by *normalizeWithinArray*s*()* function)
> for the dye effect? If this is the case, why is a dye effect term
The loess normalization done by normalizeWithinArrays() accounts for a
global dye effect trend. However it is possible that some of the
on the array might show special dye effects specific to those probes
don't follow the overall dye effect trend. It is the purpose of a dye
effect term in the linear model to allow for the possibility of such
probe-specific dye effects.
> Also, with a view to identifying differentially expressed genes, is
> sample code provided in my previous mail otherwise correct? Are
> any alterations that I should consider?
is not needed, and is obviously superfluous in your code anyway.
> On Sat, Jan 4, 2014 at 2:57 PM, Ryan <rct at="" thompsonclan.org="">
>> Hi Joseph,
>> You cannot include a dye effect term in this design, because the
>> biological effect and dye effect are completely confounded due to
>> of dye swaps. Hence, I believe this design is incapable of
>> between dye effects and biological effects. The only way to proceed
>> be to make an arbitrary assumption about the dye effects (e.g.
>> effects are zero).
>> On Sat Jan 4 09:43:02 2014, Joseph Shaw [guest] wrote:
>>> Hi all,
>>> I'm currently looking at data collected from a two-channel
>>> experiment; the experimental design is as follows:
>>> - The data represents the results of a competitive hybridization
>>> process between control RNA and treatment RNA.
>>> - The data comprises n*m slides (*n* biological replicates and
>>> technical replicates for each biological replicate).
>>> - The control label dye (cy5) treatment label dye (cy3) remain
>>> across all slides - hence, **there is no dye-swap aspect to the
>>> - The data were generated by ScanArray Express and slide data
>>> stored in separate .csv files.
>>> I'm very new to the limma package. Is it possible to use the limma
>>> package to identify differentially expressed genes for this
>>> If so,
>>> - how can the design matrix be specified? will a "dye effect"
>>> still be required even if there is no dye-swap?
>>> - is a contrast matrix necessary for this procedure?
>>> - are there any specialist normalisation techniques required for
>>> My code so far is as follows:
>>>> # Assuming the contents of the targets file have been identified:
>>>> RG<-read.maimages(targets, source="scanarrayexpress", sep=",")
>>>> RGbk <- backgroundCorrect(RG, method="normexp", offset=50)
>>>> MA <- normalizeWithinArrays(RGbk, method="loess")
>>>> MA.b=normalizeBetweenArrays(MA, method="quantile")
>>>> design <- modelMatrix(targets, ref="control") # nmx1 matrix; all
>>>> elements set to -1.
>>>> fit <- lmFit(MA, design)
>>>> fit <- eBayes(fit)
>>>> topTable(fit, coef=1, adjust="fdr")
>>> Any assistance with the above would be greatly appreciated.
>>> -- output of sessionInfo():
>>> R version 3.0.2 (2013-09-25)
>>> Platform: x86_64-apple-darwin10.8.0 (64-bit)
>>>  en_IE.UTF-8/en_IE.UTF-8/en_IE.UTF-8/C/en_IE.UTF-8/en_IE.UTF-8
>>> attached base packages:
>>>  stats graphics grDevices utils datasets methods
>>> other attached packages:
>>>  limma_3.18.7
>>> Sent via the guest posting facility at bioconductor.org.
The information in this email is confidential and