Hello,

I apologize if this question was asked before but I did not find an answer to my issue.

I would like to subset my expressionSet choosing the probes only which have gene symbol.

eSet object:

ExpressionSet (storageMode: lockedEnvironment)
assayData: 35556 features, 6 samples 
  element names: exprs 
protocolData: none
phenoData
  sampleNames: GSM1153102 GSM1153103 ... GSM1153107 (6 total)
  varLabels: title geo_accession ... data_row_count (36 total)
  varMetadata: labelDescription
featureData
  featureNames: 10338001 10338002 ... 10608724 (35556 total)
  fvarLabels: ID Symbol Name Ensembl
  fvarMetadata: Column Description labelDescription
experimentData: use 'experimentData(object)'
Annotation: mogene11sttranscriptcluster.db  

Thank you for your help.

Anita 

What do you mean by the question Gene Symbol ? Do you mean excude probesets that dont match to ENTREZID or there are not annotated ? If so, you could check the genefilter package and the function nsFilter()

Then, you could write something like below:

eset <- nsFilter(eSet, require.entrez=TRUE, remove.dupEntrez=FALSE, var.filter=FALSE, var.cutoff=0.5, feature.exclude="^AFFX")$eset

where you remove probesets without an Entrez Gene ID annotation. 

Hi,

Thanks for your quick reply. Yes, I meant annotation : probesets with Symbol or  Entrezid.  Thanks for the code and the package suggestion. I will look into it.

Thanks once again, 

Anita