[BioC] how to make an MA plot with single color agilent microarray?

Gordon K Smyth smyth at wehi.EDU.AU
Sun Jun 26 03:50:26 CEST 2011


Dear Richard and Wolfgang,

The strategy of plotting relative to a reference array that Wolfgang 
mentions, is what limma does automatically for single channel data.

Best wishes
Gordon

> Date: Sat, 25 Jun 2011 07:52:20 +0000
> From: Richard <greener at uw.edu>
> To: <bioconductor at stat.math.ethz.ch>
> Subject: Re: [BioC] how to make an MA plot with single color agilent
> 	microarray?
>
> Wolfgang Huber <whuber at ...> writes:
>
>>
>> Hi Richard
>>
>> Il Jun/19/11 4:19 AM, Richard Green ha scritto:
>>> Could someone suggest the best way to create an MA plot from 14 single color
>>> agilent arrays? I would really appreciate it. Thanks -Rich
>>>
>>
>> One option is to compute a virtual mean (or median) reference array and
>> to make 14 MA-plots of each array against that reference. Essentially,
>> you redefine the 'A' to be the average of all arrays rather than just of
>> a pair, and then compute 14 different 'M's.
>>
>> The MA plot is just an orthogonal transformation of the ordinary 2D
>> scatterplot. In that sense, any other linear transformation of your
>> 14-dimensional dataset could also be useful (e.g. PCA), in particular if
>> one of the new axes aligns with the (1,1,1,..,1) direction.
>>
>> (A variant of the 'virtual reference array' idea is used in the MA plots
>> in the reports generated by 'arrayQualityMetrics', which you can see by
>> running the R script associated with the vignette
>> http://www.bioconductor.org/packages/devel/bioc/html/arrayQualityMetrics.html
>>   -> "Introduction: microarray quality assessment with arrayQualityMetrics")
>>
>> Best wishes
>> Wolfgang Huber
>> EMBL
>> http://www.embl.de/research/units/genome_biology/huber
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at ...
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives:
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>
>>
> Hi Wolfgang, Thanks for your response. I tried to install arrayQualityMetrics
> but was unable (see errors below). Not sure why, tried it on a couple of
> different R installations. I did figure out how to create the individual MA
> plots but am now just trying to figure out the simplest way to overlay two MA
> plots on to each other to see how different they look. I have generated a PCA
> but would like to compare pairs of arrays by MA plot. Any suggestion are
> appreciated. I pasted my R script below:
> source("http://www.bioconductor.org/biocLite.R")
> library(limma)
> setwd("/vol04/microarray/")
> RG <-
> read.maimages(files=dir(),source="agilent",columns=list(G="gMeanSignal",Gb="gBGM
> edianSignal",R="gMeanSignal",Rb="gBGMedianSignal"))
> RG <- backgroundCorrect(RG, method="normexp", offset=50)
> RG <- normalizeBetweenArrays(RG$R, method="quantile")
> RG <- log2(RG)
> for (i in 1:24) {
> fname<-paste("/vol04/microarray/ma_plot",i,".png",sep="")
> png(fname,300,300)
> print(plotMA(RG, array=i, cex=.2, ylim=c(-1.5,1.5)))
> dev.off()
> }
>
> The above works no problem but when I try to add more than one array I get
>
>> print(plotMA(RG, array=(1:3,13:16), cex=.2, ylim=c(-1.5,1.5)))
> Error: unexpected ',' in "print(plotMA(RG, array=(1:3,"
>> print(plotMA(RG, array=1:13, cex=.2, ylim=c(-1.5,1.5)))
> Error in xy.coords(x, y, xlabel, ylabel, log) :
>  'x' and 'y' lengths differ
>
>
> ###############quality metrics error ###############################
> ERROR: configuration failed for package ?XML?
> * removing ?/home/greener/R/x86_64-redhat-linux-gnu-library/2.12/XML?
> ERROR: dependency ?XML? is not available for package ?SVGAnnotation?
> * removing ?/home/greener/R/x86_64-redhat-linux-gnu-library/2.12/SVGAnnotation?
> ERROR: dependencies ?SVGAnnotation?, ?XML? are not available for package
> ?arrayQualityMetrics?
> * removing ?/home/greener/R/x86_64-redhat-linux-gnu-
> library/2.12/arrayQualityMetrics?

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



More information about the Bioconductor mailing list