[BioC] Problem with NChannelSet objects using Agilent Data with arrayQualityMetrics package

[Wolfgang Huber]

Dear Laurent
try with this script:

----------------------
library("ArrayExpress")
library("arrayQualityMetrics")
library("geneplotter")
library("vsn")

AEset = ArrayExpress("E-GEOD-32915")

## explore raw data distributions
par(mfrow=c(2,1))
for(w in c("E", "Eb")) multidensity(log2(assayData(AEset)[[w]]), main=w)

## normalize
v = vsn2(assayData(AEset)$E - assayData(AEset)$Eb)

x = ExpressionSet(assayData = exprs(v),
    phenoData = phenoData(AEset),
    featureData = featureData(AEset),
    experimentData = experimentData(AEset))

arrayQualityMetrics ( x, intgroup = "Comment..Sample_source_name." )
-------------------------

Usually, as is here, some data wrangling and “normalization” is required between downloading a raw data set from ArrayExpress and processing it further with Bioconductor tools. Although the microarray field has been around for a while, these steps are still not fully automated. (Contributions are welcome.) The code above implies the (subjective) decision to subtract Eb from E, and then call vsn2 for between-array “normalization” and log-like transformation. I then stick the result back into ExpressionSet ‘x’ and so we go.

Hope this helps
	Wolfgang

> print(sessionInfo())
R Under development (unstable) (2014-02-27 r65088)
Platform: x86_64-apple-darwin13.0.0/x86_64 (64-bit)

locale:
[1] C

attached base packages:
[1] parallel  stats     graphics  grDevices utils     datasets  methods  
[8] base     

other attached packages:
 [1] vsn_3.31.2                 geneplotter_1.41.1        
 [3] annotate_1.41.2            AnnotationDbi_1.25.9      
 [5] lattice_0.20-27            arrayQualityMetrics_3.19.6
 [7] ArrayExpress_1.23.1        Biobase_2.23.5            
 [9] BiocGenerics_0.9.3         fortunes_1.5-2            

loaded via a namespace (and not attached):
 [1] BeadDataPackR_1.15.1  BiocInstaller_1.13.3  Biostrings_2.31.14   
 [4] Cairo_1.5-5           DBI_0.2-7             Formula_1.1-1        
 [7] Hmisc_3.14-1          IRanges_1.21.32       KernSmooth_2.23-10   
[10] RColorBrewer_1.0-5    RJSONIO_1.0-3         RSQLite_0.11.4       
[13] Rcpp_0.11.0           SVGAnnotation_0.93-1  XML_3.98-1.1         
[16] XVector_0.3.7         affy_1.41.3           affyPLM_1.39.2       
[19] affyio_1.31.0         beadarray_2.13.1      cluster_1.14.4       
[22] colorspace_1.2-4      gcrma_2.35.1          genefilter_1.45.1    
[25] grid_3.1.0            gridSVG_1.4-0         hwriter_1.3          
[28] latticeExtra_0.6-26   limma_3.19.21         plyr_1.8.1           
[31] preprocessCore_1.25.5 reshape2_1.2.2        setRNG_2011.11-2     
[34] splines_3.1.0         stats4_3.1.0          stringr_0.6.2        
[37] survival_2.37-7       xtable_1.7-1          zlibbioc_1.9.0 


Il giorno 26 Feb 2014, alle ore 17:43, Laurent.Pays <Laurent.Pays at univ-lyon1.fr> ha scritto:

> Hi to all,
> 
> After downloading Agilent Data  using the ArrayExpress package:
> 
>>> AEset <-ArrayExpress("E-GEOD-32915”)
> 
> I verified that AEset was indeed a NChannelSet object:
>>> class(AEset)
>> [1] "NChannelSet"
>> attr(,"package")
>> [1] “Biobase"
> 
> Then I tried to use the arrayQualityMetrics package which should work on NChannelSet objects:
> 
>> arrayQualityMetrics ( AEset )
> 
> But I get this error message:
> 
>> The report will be written into directory 'arrayQualityMetrics report for AEset'. 
>> Erreur dans platformspecific(expressionset, intgroup, do.logtransform) : 
>>  'assayData(expressionset)' must contain either 'exprs', or 'R' and 'G', but not both.
> 
> My question is how do I get pass this error, or at least or can I remove either ‘exprs’, or ‘R’ and ‘G’ from the NChannelSet object knowing that when I try:
>>> channelNames(AEset)
> I only get:
>> [1] "E"  “Eb”
> and I don’t know what to analyse with arrayQualityMetrics: “E” or “Eb”?
> 
> Thanks in advance to your answers
> 
> Laurent
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor