[BioC] Importing quantile normalized .txt file to limma for differential expression

James W. MacDonald jmacdon at uw.edu
Wed Sep 25 15:59:00 CEST 2013


Hi Dhanjit,

Please don't take conversations off list (e.g., use 'Reply-all' when 
responding). We like to think of the archives as a repository of 
knowledge, and if you make things private, then that use is diminished.

I am not sure where you got the procedure you are using for processing 
the data, but you are doing it all wrong. There is no need to repeatedly 
write data to disk, only to read it back in again. In addition, as the 
warning states, you should really be using either oligo or xps to 
process your data. Additionally, you are only quantile normalizing the 
probes without then summarizing them. Unless you summarize the data, you 
are not getting a good measure of expression (plus you now have massive 
duplication of the measurements for each transcript).

A quick start primer for this would be (after installing oligo - if you 
want to use xps, please see the vignettes found here: 
http://www.bioconductor.org/packages/release/bioc/html/xps.html).

library(oligo)
dat <- read.celfiles(list.celfiles())
eset <- rma(dat)

This gives you summarized, normalized, background corrected data. You 
can then feed the eset object directly into limma, without writing it to 
disk. You first need a design matrix. I will make assumptions here about 
the structure of your data, but note that this will almost certainly not 
be correct for what you have in hand. You will have to modify to suit.

samps <- factor(rep(c("control","sample"), each = 8))
design <- model.matrix(~samps)
fit <- lmFit(eset, design)
fit2 <- eBayes(fit)
topTable(fit2, coef=2)

And that's it! You can now output in text files, HTML tables, whatever 
you like. You should certainly peruse the limma User's Guide to get more 
detailed information, and you should also look at things like annaffy or 
ReportingTools, which make it simple to output annotated data in both 
HTML and text format.

Best,

Jim




On 9/25/2013 9:18 AM, Dr. Dhanjit Kumar Das wrote:
> Dear Sir,
>
> Thank you so much for your prompt reply.
>
> I would like to give you more details about my study. I run 16 samples 
> comprising of two groups (8 controls & 8 samples) in affy Gene1.0st 
> arrays. I have done trmean function and carried out quantile 
> normalization using R. I am attaching the R script as PDF for your 
> reference. After quantile normalization, the normalized expression 
> values table was written using script "write.table(exprSet.quantile, 
> file="C:/Dhanjit/IUGR/IUGR-FINAL/IUGR-exrSet.quantile1.txt", quote=F, 
> sep="\t")". These normalized values were saved as expression set 
> (exprSet.quantile) and this expression set needs to be read by limma. 
> When the matrix is designed using exprSet.quantile, only one samples 
> has been shown (Result was shown at the attached file).
>
> I have the queries how to read this expression set (exprSet.quantile) 
> for designing the matrix for differential expression using limma.
>
> Kindly suggest me the procedure for analysis of differential expression.
>
> Thanking you
>
> Dhanjit
> _____________________________________________
> */Dr. Dhanjit Kumar Das/*, Ph.D
> Scientist 'B'
> Genetic Research Centre
> National Institute for Reserach in Reproductive Health
> Jehangir Merwanji Street
> Parel, Mumbai-400 012
> INDIA
> Phone: +91-22-24192108 / 2145
> Fax:   +91-22-24147843
> *____________________________________*
>
> ------------------------------------------------------------------------
> *From:* James W. MacDonald <jmacdon at uw.edu>
> *To:* Dhanjit [guest] <guest at bioconductor.org>
> *Cc:* bioconductor at r-project.org; dkdas_grc at yahoo.com
> *Sent:* Tuesday, September 24, 2013 8:35 PM
> *Subject:* Re: [BioC] Importing quantile normalized .txt file to limma 
> for differential expression
>
> You will need to give more information than this if you want help. How
> are you pre-processing the data? Using R or something else? What have
> you tried, and what were the results? Note that you cannot import
> anything into limma - you import into R and then use limma to analyze
> the data.
>
> Also, have you read the limma User's Guide? If not, you should, as
> probably 99.325% (just an estimate) of your questions are covered in
> that document.
>
> http://www.bioconductor.org/packages/2.12/bioc/vignettes/limma/inst/doc/usersguide.pdf
>
> Best,
>
> Jim
>
>
>
> On Tuesday, September 24, 2013 9:04:07 AM, Dhanjit [guest] wrote:
> >
> > I am using limma package for identifying differential expression of 
> affymetrix .CEL file. I used quantile normalization method and saved 
> the normalized data in .txt file. However, i am not able to import the 
> .txt file into limma for differential expression.
> >
> > Kindly tell me the sequential procedure for identifying the 
> differential expression of quantile normalized data.
> >
> >  -- output of sessionInfo():
> >
> > aa
> >
> > --
> > Sent via the guest posting facility at bioconductor.org.
> >
> > _______________________________________________
> > Bioconductor mailing list
> > Bioconductor at r-project.org <mailto:Bioconductor at r-project.org>
> > https://stat.ethz.ch/mailman/listinfo/bioconductor
> > Search the archives: 
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
> --
> James W. MacDonald, M.S.
> Biostatistician
> University of Washington
> Environmental and Occupational Health Sciences
> 4225 Roosevelt Way NE, # 100
> Seattle WA 98105-6099
>
>

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099



More information about the Bioconductor mailing list