This is a test version of Biostars. For the public version, visit https://www.biostars.org.
variance stabilization and normalization

Hi,

I am going to perform variance stabilization on my data sets then I did like below

library(affy)
library(vsn)

celFiles <- list.celfiles()
affyraw=ReadAffy(filenames = celFiles)
fit = vsn2(affyraw)
vsn2: 506944 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.

nkid = predict(fit, newdata=affyraw)
vsnrma(nkid)
vsn2: 506944 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.
Creating a generic function for ‘nchar’ from package ‘base’ in package ‘S4Vectors’

Calculating Expression
ExpressionSet (storageMode: lockedEnvironment)
assayData: 22810 features, 8 samples 
  element names: exprs 
protocolData
  sampleNames: Col-0 24h primed.CEL.CEL Col-0 24h unprimed.CEL.CEL ... Col-0 8h unprimed.CEL.CEL (8 total)
  varLabels: ScanDate
  varMetadata: labelDescription
phenoData
  sampleNames: Col-0 24h primed.CEL.CEL Col-0 24h unprimed.CEL.CEL ... Col-0 8h unprimed.CEL.CEL (8 total)
  varLabels: sample
  varMetadata: labelDescription
featureData: none
experimentData: use 'experimentData(object)'
Annotation: ath1121501 

a <- vsnrma(nkid)
vsn2: 506944 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.
Calculating Expression

celFiles <- list.celfiles()
affyraw=ReadAffy(filenames = celFiles)
eset <- justvsn(celFiles)
Error in (function (classes, fdef, mtable)  : 
  unable to find an inherited method for function ‘vsn2’ for signature ‘"character"’
justvsn(a)
vsn2: 22810 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.
ExpressionSet (storageMode: lockedEnvironment)
assayData: 22810 features, 8 samples 
  element names: exprs 
protocolData
  sampleNames: Col-0 24h primed.CEL.CEL Col-0 24h unprimed.CEL.CEL ... Col-0 8h unprimed.CEL.CEL (8 total)
  varLabels: ScanDate
  varMetadata: labelDescription
phenoData
  sampleNames: Col-0 24h primed.CEL.CEL Col-0 24h unprimed.CEL.CEL ... Col-0 8h unprimed.CEL.CEL (8 total)
  varLabels: sample
  varMetadata: labelDescription
featureData: none
experimentData: use 'experimentData(object)'
Annotation: ath1121501 

a <justvsn(a)
vsn2: 22810 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.
Error in a < justvsn(a) : 
  comparison (3) is possible only for atomic and list types

a <- justvsn(a)
vsn2: 22810 x 8 matrix (1 stratum). Please use 'meanSdPlot' to verify the fit.
r
• 5,321 views • link
updated by Ram Ram • written by zizigolu zizigolu

4 answers

bg.correct should stand for background correct = FALSE

You read your CEL files with the ReadAffy function, and immediately apply the normalization without substracting the background. I don't think this is correct

The other attributes apply different methods, including the normalize.method where you are indicating the vsn method for normalization.

I don't know the expresso function, but it seems that it includes the possibility of doing a background substraction. Read the vignette with a ?expresso and it will clarify to you what it can do

• 0 views • link
updated by Ram Ram • written by Antonio R. Franco Antonio R. Franco

may you please help me to write the correct?

• 0 views • link
written by zizigolu zizigolu

You need to investigate the expreso function since bg.correct can mean either of two things. That a correction has not been done, therefore it must be done, or that a background correction is optative

Beyond that the code included in the vignette is the one you must follow

• 0 views • link
updated by Ram Ram • written by Antonio R. Franco Antonio R. Franco

After loading the affy and vsn libraries..

Into R run

?expresso

You can read information about the function. In this case, this is written

bg.correct
a boolean to express whether background correction is wanted or not.

In your case, you need to run a bg.correct= TRUE

• 0 views • link
updated by Ram Ram • written by Antonio R. Franco Antonio R. Franco

thank you so much Antonio

• 0 views • link
updated by Ram Ram • written by zizigolu zizigolu

Hi,

I collected this code finally

# performing variance stabilization and normalization on microarray data sets
# loading required libraries
library(affy)
library(vsn)
# listing the cel files
celFiles <- list.celfiles()
# assigning the cel files to affyraw variable
affyraw=ReadAffy(filenames = celFiles)
# performing vsn normalization
vsn.data <- expresso(affyraw, normalize.method="vsn", bg.correct=F, pmcorrect.method="pmonly", summary.method="medianpolish")
# examining the normalization
boxplot(affyraw,col="red")
plot(exprs(affyraw)[,1:2], log = "xy", pch=".",
     main="all")
# writing the result
write.table(vsn.data, file = "vsn.txt", dec = ".", sep = "\t", quote = FALSE)
• 0 views • link
updated by Ram Ram • written by zizigolu zizigolu

thank you Fereshteh, you are amazing!!!

• 0 views • link
written by zizigolu zizigolu

Antonio,

Actually I did some sporadic code from internet I can't get what they mean, even finally I could not get AGI IDs and the result is with prob IDs

In brief I don't know 

vsn.data <- expresso(affyraw, normalize.method="vsn", bg.correct=F, pmcorrect.method="pmonly", summary.method="medianpolish")

what means....

• 0 views • link
updated by Ram Ram • written by zizigolu zizigolu

Log in to answer this question.

More posts like this