Hi,

I'm new to R and genetics and I am trying to processes some gene expression data. So I can download some cel.gz files and then I can put them in a pipeline to get a gene expression file for that cel.gz file. I am using an high performance computing cluster so it isn't too easy to install different versions of R or bioconductor (especially newer ones), so I am working with our latest versions which are R/3.3.2 and Bioconductor 3.4.

So say I have a file GSM3921.cel.gz. In my script I do:

library(affy)
library(frma)
abatch <- ReadAffy(filenames="GSM3921.cel.gz")
aobj <- frma(abatch)
aeset <- exprs(aobj)

This gives me the expression values for the probes. However, what I really want to do is convert these probe values to entrez values. It looks like once upon a time ago you could simply do:

library(affy)
library(frma)
library(hgu133afrmavecs)
library(hgu133ahsentrezgcdf)
abatch <- ReadAffy(filenames="GSM3921.cel.gz",cdfname="HGU133A_HS_ENTREZG") 
aobj <- frma(abatch,input.vecs=hgu133ahsentrezgfrmavecs)
aeset <- exprs(aobj)

However, when I try to install library(hgu133ahsentrezgcdf) I get an error saying this package is not available for R/3.3.2. I can't really find any documentation on this package. Is there a way to overcome this installation problem, or is there any other way to convert the expression values to entrez? I also need this to work in a similar way for HGU133PLUS2 files as well.

One more note, is that when reading posts about this is seems like anyone using library(hgu133ahsentrezgcdf) is doing so with R/3.0.2 or older. I tried downloading this library with R/3.0.1 and I got the same this package is not available for R/3.3.2 error, so maybe I'm not downloading it correctly? To download it I do:

source("http://bioconductor.org/biocLite.R")
biocLite()
biocLite("hgu133ahsentrezgcdf")

Any help getting this entrez conversion figured out would be great.

Unfortunately biocLite() will not automagically download the custom (remapped) CDF you would like to use from the repository of the MBNI group. You rather have to download them directly by:

install.packages("http://mbni.org/customcdf/22.0.0/entrezg.download/hgu133ahsentrezgcdf_22.0.0.tar.gz", type="source", repos=NULL)
install.packages("http://mbni.org/customcdf/22.0.0/entrezg.download/hgu133ahsentrezgprobe_22.0.0.tar.gz", type="source", repos=NULL)
install.packages("http://mbni.org/customcdf/22.0.0/entrezg.download/hgu133ahsentrezg.db_22.0.0.tar.gz", type="source", repos=NULL)

This will download and install the latest version (= v22) of the custom CDF for the HGU133a array that is based on the Entrez gene database.

Also, your line of code in which you specify the use of the custom CDF is not fully OK; it should be:

abatch <- ReadAffy(filenames="GSM3921.cel.gz", cdfname="hgu133ahsentrezg")

Please note that I don't know whether you can actually use the frma vectors with a custom/remapped CDF.

You can use the entrez gene remapped CDFs with frma; however, I’m several versions behind the Brain Array folks. If you look at the version field in the frozen parameter vector object, it will tell you the corresponding alternative CDF version to use. 

Thanks Guido and Matthew, your responses worked perfectly! I had to install (= v19) of the custom cdf as this is what verison worked with frma.

As a small side note, Matthew in your DESCRIPTION for the package hgu133plus2hsentrezgfrmavecs, it says the package was created with hgu133ahsentrezgcdf, and I think it should say hgu133plus2hsentrezgcdf.