How to annotate GEO microarray datasets with GEOquery?

Question

[grayapply2009]

Hi, I read the GSE file into R as follows.

gse <- getGEO("GSE4928", GSEMatrix=TRUE)

Now I want to convert all probe IDs to gene symbols and write the entire annotated dataset back to my computer. What should I do? By the way, which dataset is stored in gse? SOFT formatted family file, MINiML formatted family file or Series Matrix File?

Answer 1

[Sean Davis]

Take a look at this code:

gse = gse[[1]] # get just the first element in the list
head(fData(gse))
symbols = fData(gse)[,'Gene Symbol']

Now, you have the gene symbols. The gse object (after choosing the first element in the list using gse[[1]]) contains the information from the Series Matrix file AND the GPL file.

[grayapply2009]

Thank you Sean. Now how can I export the Series Matrix file with the probes replaced by gene symbols?

[Ce Gao]

After that, you can use:

expr_mat = exprs(gse)        # get the expression matrix
rownames(expr_mat) = symbols # Annotate the row names with gene symbols

[grayapply2009]

Thank you, gaoce. That works perfectly. Where can I get the tutorial for this? The GEOquery document doesn't seem to have what you guys showed me above.

[Sean Davis]

Unfortunately, this solution is not a general one due to the fact that not all GEO series have gene symbol or other annotation. What that means is that it is best to understand why the steps above work so that when you have another data set, you can follow the logic to come up with your own solution. That said, I am always interested in documentation improvements, so it you'd like to contribute, GEOquery is on github where I can accept pull requests.

[Amir]

Hi, can you help me with a similar problem? I used the code Sean put here and I change the series number to ones from the data that I want to use. But I face an error:

> symbols = fData(gse)[,'Gene Symbol']
Error in `[.data.frame`(fData(gse), , "Gene Symbol") : 
  undefined columns selected

I looked in the fData(gse) corresponding to my own GSE file and there isn't a Gene Symbol column.

ID of the dataset that I'm using is GSE65106 Could you possibly be able to help me? I would really appreciate it. Best regards