Question: Extract Pre-Normalisation Expression Data From Probe Set Values
0
gravatar for munch
5.5 years ago by
munch300
Munich
munch300 wrote:

I want to take a look on the pre-normalised data of some Affymetrix .CEL files (HuGene-1_0-st-v1, so no MM spots on the chip). For this I have to combined/summarised low level probe intensity values into probe sets values (that map to genes). For comparison I provide the first line of the Expression set:

(0) my normal approach to normalise the data:

library(affy)  
celFiles <- list.celfiles(path_to_cel_files,full.names=TRUE)
affyExpressionFS <- read.celfiles(celFiles)
eset <- rma(affyExpressionFS)
> exprs(eset)[1,]
100712_KMT_1-1_(HuGene-1_0-st-v1).CEL 120712_KMT_1-2_(HuGene-1_0-st-v1).CEL 
                             7.573983                              6.960390 
120712_KMT_1-3_(HuGene-1_0-st-v1).CEL  120712_KMT_M1_(HuGene-1_0-st-v1).CEL 
                             7.884465                              7.956841 
 120712_KMT_M2_(HuGene-1_0-st-v1).CEL  120712_KMT_M3_(HuGene-1_0-st-v1).CEL 
                             7.825583                              8.862901

(1) without normalisation using justRMA:

eset <- justRMA(background=FALSE, normalize=FALSE)
> exprs(eset)[1,]
100712_KMT_1-1_(HuGene-1_0-st-v1).CEL 120712_KMT_1-2_(HuGene-1_0-st-v1).CEL 
                             8.135738                              7.349999 
120712_KMT_1-3_(HuGene-1_0-st-v1).CEL  120712_KMT_M1_(HuGene-1_0-st-v1).CEL 
                             7.767440                              8.251783 
 120712_KMT_M2_(HuGene-1_0-st-v1).CEL  120712_KMT_M3_(HuGene-1_0-st-v1).CEL 
                             7.975933                              8.804127

(2) without normalisation using expresso:

affy.data <- ReadAffy()
eset <- expresso(affy.data, bg.correct=FALSE, normalize=FALSE, summary.method = "avgdiff", pmcorrect.method="pmonly", bgcorrect.method = "rma")
> exprs(eset)[1,]
100712_KMT_1-1_(HuGene-1_0-st-v1).CEL 120712_KMT_1-2_(HuGene-1_0-st-v1).CEL 
                             7.564649                              6.946216 
120712_KMT_1-3_(HuGene-1_0-st-v1).CEL  120712_KMT_M1_(HuGene-1_0-st-v1).CEL 
                             7.875625                              7.947773 
 120712_KMT_M2_(HuGene-1_0-st-v1).CEL  120712_KMT_M3_(HuGene-1_0-st-v1).CEL 
                             7.826963                              8.859935

(3) with normalisation using justRMA:

eset <- justRMA(background=TRUE, normalize=TRUE)
> exprs(eset)[1,]
100712_KMT_1-1_(HuGene-1_0-st-v1).CEL 120712_KMT_1-2_(HuGene-1_0-st-v1).CEL 
                             7.564649                              6.946216 
120712_KMT_1-3_(HuGene-1_0-st-v1).CEL  120712_KMT_M1_(HuGene-1_0-st-v1).CEL 
                             7.875625                              7.947773 
 120712_KMT_M2_(HuGene-1_0-st-v1).CEL  120712_KMT_M3_(HuGene-1_0-st-v1).CEL 
                             7.826963                              8.859935

My conclusion is that (0),(2),(3) is the same. My question now is: Is (1) the best method to get the pre-normalised probe sets values? I am right that the expresso function in (2) perform normalisation although I use normalize=FALSE? Iam right that the values in (1) are also log2 values? Thank you!

ADD COMMENTlink modified 5.5 years ago by Johanna Schott380 • written 5.5 years ago by munch300
1
gravatar for Johanna Schott
5.5 years ago by
Germany
Johanna Schott380 wrote:

Due to my very limited experience with the affy package I won't directly answer your questions. But I would like to point out that the oligo package is an alternative that you could compare with your approach (1). What I would do to get summarized but not normalized or background corrected (log2 transformed) data is:

library(oligo) celFiles <- list.celfiles(path_to_cel_files, full.name = TRUE) data <- read.celfiles(celFiles) eset <- rma(data, object, background = FALSE, normalize = FALSE) exprs(eset)[1,]

rma() with the oligo package will perform median polish for probe summarization.

ADD COMMENTlink written 5.5 years ago by Johanna Schott380

Thank you, it is the same result as (1), so I think I can rely on the correctness.

ADD REPLYlink modified 5.5 years ago • written 5.5 years ago by munch300
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1136 users visited in the last hour