RMA normaliztion using XPS
10
0
Entering edit mode
@dhwanidholakia-7550
Last seen 2.4 years ago
INDIA

hey

Can any one help me with how to normalize 5000 .CEL files using XPS. The steps given to root schema seems very complicated

normalization xps • 2.1k views
2
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

1, Why do you give file names starting with 'Test3'. Please note that 'Test3' was an Affymetrix test array,
which I have used to to create the demo files. Do you currently use the Test3 array or do you work with
your CEL-files from the HT_HG-U133A array?

In the later case you should not call your scheme 'scheme.test3' but e.g. 'scheme.hthgu133a' as I have shown
you in the script 'xps/examples/script4schemes.R' (lines 352-356).

2, I would suggest that you create one directory for your CEL-files and another directory for your results.
Thus your code would look as follows:

celdir <- "/home/dhwani/Desktop/cmap_data/final_cmap_filtering/array_specific/ht_all/raw/"
datdir <- "/home/dhwani/Desktop/cmap_data/final_cmap_filtering/array_specific/ht_all/results/"

scheme.hthgu133a <- root.scheme("/home/dhwani/Desktop/xps_files/schemes/SchemeHTHGU133A.root")
data.hthgu133a <- import.data(scheme.hthgu133a, "DataHTHGU133A", filedir=datdir, celdir=celdir)

data.bg.rma <- bgcorrect.rma(data.hthgu133a,"HTHGU133ARMABgrd", filedir=datdir)
data.qu.rma <- normalize.quantiles(data.bg.rma, "HTHGU133ARMANorm", filedir=datdir)
data.mp.rma <- summarize.rma(data.qu.rma, "HTHGU133ARMAExpr", filedir=datdir, tmpdir="")
expr.rma <- validData(data.mp.rma)

3, In my reply to you 2 days ago I have written:
When running function 'rma()' your values are normalised, however, as mentioned on page 8 of vignette 'xps.pdf'
and in the help '?rma':
Note: In contrary to other implementations of RMA the expression measure is given to you in linear scale,...

This means that you have to convert the expression data to log2 values, i.e.

expr.rma  <- validData(data.mp.rma)
expr.rma2 <- as.data.frame(log2(expr.rma))

4, From your sessionInfo() I see that you have 'xps_1.20.3', which is a very old version. Please update to

0
Entering edit mode

Thank you so much sir for all your information finally i am able to use xps smoothly however probes names vary when normalised using affy package

1
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

hey

The steps given to root schema are explained for most arrays step-by-step in:

xps/examples/script4schemes.R.

You can simply copy/paste the example code and change the directories!

The steps how to normalize CEL files are explained step-by-step in:

xps/examples/script4xps.R

xps/examples/script4exon.R

xps/examples/script4xpsPreprocess.R

For 5000 CEL files I would do the normalization stepwise so that you need not start from beginning if something goes wrong. Furthermore, test everything first using only 10 CEL files so that you know that it works in principle.

Examples how to do stepwise normalization are shown in:

xps/examples/script4exon.R

xps/examples/script4xpsPreprocess.R

You did not mention which array you want to normalise. Furthermore, you did not gove your sessionInfo()

Christian

0
Entering edit mode

thank you so much sir

i tried the script but i am not able to understand i i have to create root schema for my rma normalization also?

1) i have .CEL files of HT_HG-U133A . i am very confused how to make scehma files and from where to get the probe files for the same as the probe file is not available in the library files only .cdf file is available

1
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

Please read vignette 'xps.pdf'. Appendix A1 explains how to create root scheme files. For every array you have to create this file only  once and then you can use it for all your work with this array.

As I have mentioned example code how to create the scheme files is shown in xps/examples/script4schemes.R

The example code for HT_HG-U133A is shown in lines 352-356.

http://www.affymetrix.com/support/technical/byproduct.affx?product=ht_hg-u133_set

0
Entering edit mode

Thank you sir , i was able to make the root schema

0
Entering edit mode

Dear sir

i was able to make root schema and then i imported .CEL files but i am unable to do rma normalization and view the normalized results

eset.rma <- expresso(Data.Tes3,bg.correct="rma",normalize.method="quantiles",pmcorrect.method="pmonly",summary.method="medianpolish")

and when i do

> data.rma <- rma(data.test3,"Test2RMA",tmpdir="")

my values are not normalized rather they are in range of more than 1 to 1000

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

From where did you get the command 'eset.rma <- expresso(Data.Tes3,...)'?

Please note that package 'xps' does neither have a function 'expresso()' nor does it support 'eset'. The corresponding command in 'xps' is 'express()', see the help '?express'. For a detailed explanation of function 'express()' see the vignette 'xpsPreprocess.pdf'.

When running function 'rma()' your values are normalised, however, as mentioned on page 8 of vignette 'xps.pdf' and in the help '?rma':

Note: 'In contrary to other implementations of RMA the expression measure is given to you in linear scale,..'

0
Entering edit mode

is the below command correct to normalize values

data.rma <- rma(data.test3,"Test2RMA",tmpdir="")

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

Yes, however, please type '?rma' to see the meaning of the different parameters of 'rma()' and some examples.

0
Entering edit mode

because of some memory reason my system stooped in between and the last file generated was Test3RMANorm.root.

Can i get normalized data frame using the above file

0
Entering edit mode

If we have uploaded the .CEL files is it possible to use that .root file for further processing

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

0
Entering edit mode

below is my session info

R version 3.1.3 (2015-03-09)
Platform: x86_64-redhat-linux-gnu (64-bit)
Running under: CentOS release 6.6 (Final)

locale:
[1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8       LC_NAME=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

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

other attached packages:
[1] xps_1.20.3

loaded via a namespace (and not attached):
[1] tools_3.1.3

0
Entering edit mode
savehistory(file = ".Rhistory") this command doesnt work to save history . what is the way i can save history so that i can so u my commands if there was some error in commands
0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

Usually, this is a question for the R-help mailing list. However, if you google 'R save history' you will find e.g.:

http://stackoverflow.com/questions/5628759/r-sometimes-does-not-save-my-history

https://nsaunders.wordpress.com/2013/12/03/r-how-not-to-use-savehistory-and-source/

0
Entering edit mode

Below is the command history of my xps rma normalization. please help me to identify the error in script as the values after normalization varies in wide range 10 - 39000

celdir <- "/home/dhwani/Desktop/cmap_data/final_cmap_filtering/array_specific/ht_all/"
scheme.test3 <- root.scheme("/home/dhwani/Desktop/xps_files/schemes/SchemeTest3.root")
data.test3 <- import.data(scheme.test3, "DataTest3", celdir=celdir)
data.bg.rma <- bgcorrect.rma(data.test3,"Test3RMABgrd",filedir=celdir)
data.qu.rma <- normalize.quantiles(data.bg.rma,"Test3RMANorm",filedir=celdir)
data.mp.rma <- summarize.rma(data.qu.rma,"Test3RMAExpr",filedir=celdir,tmpdir="")
expr.rma <- validData(data.mp.rma)

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

ad 1) What do you mean with 'probe names vary'?
Probeset names in xps are identical to the ones shown in the Affymetrix annotation files, which you can open
with Excel or any text editor to prove it yourself. However, it may depend on the version of the annotation
files that you have used. The current version is na34. I do not know which version you are using for xps and
which version you are using for affy.

as 2) Unless you compile xps yourself, the version of xps to use depends on the Bioconductor version you are using.
For updating Bioconductor packages please see:
http://www.bioconductor.org/install/

0
Entering edit mode

after your prevuous comments i updated xps to 1.22 version and or annotation i am using na35

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

xps_1.22 is still an old version. Your sessionInfo() says that you are using R version 3.1.3 (2015-03-09),
which is used by Bioconductor version 3.0, and this contains xps_1.26.1.

BTW, just this weekend both R and Bioconductor were updated to new versions (R 3.2.0 and BioC 3.1) with xps_1.28.0

0
Entering edit mode
cstrato ★ 3.9k
@cstrato-908
Last seen 3.3 years ago
Austria

User dhwani.dholakia wrote Comment: RMA normaliztion using XPS:

After normalaization my file names completly changed and now it becomes difficult for further analysis.
what can  i do get row names as orignal file names

First, is there a reason why I do not see your question on support.bioconductor.org?

Second, without an example of your code and  output I do not know what you mean. So please show me your code
and some lines of the output you get, e.g. something like:

expr.rma2 <- as.data.frame(log2(expr.rma))