Combining Synonymous Affymetrix Probesets (For One Gene)

Question

[Mikhail Pyatnitskiy]

What are common ways to combine several values from synonymous Affy probesets (relating to one gene)? Any literature or links? Thanks in advance.

Answer 1

[User 59]

Already covered in detail here: R + Bioconductor : Combining Probesets In An Expressionset

You might also be interested in these two threads:

• Is It Possible For Two Different Affymetrix Probe Set Id To Have Common Annotations To Same Gene ?
• Microarray Gene Annotation

Answer 2

[Will]

My solution has been to use the "BrainArray" custom CDFs. These are created, and updated weekly, to reconstruct the affy-probesets so that each probeset matches a SINGLE ID. They have a build for UniGene (every probeset matches to a single UniGene ID), Entrez Gene, Entrez Protein, and dozens of others.

I've found that this make my downstream annotation much easier when I'm dealing with gene and protein level annotations. The only problem is that you need to have the RAW CEL data to use these CDFs.

Hope that helps, Will

[Khader Shameer]

Are you sure it is being updated regularly ? I could see the last update was on 2009. http://brainarray.mbni.med.umich.edu/Brainarray/Database/CustomCDF/CDF_download.asp#v12

[User 59]

Khader the mailing list is still active with the maintainers participating. I don't believe it's ever been updated weekly!

[Will]

maybe i just misread the "weekly" part somewhere. I do know that they are on the most recent build of the human genome at least.

[Khader Shameer]

Thanks guys for clarifying this.

Answer 3

[Yuri]

I agree that with custom CDF you probably can summarize different probesets on gene level. However, sometime you have access only to processed data on probeset level, not to raw data (CEL files). Or just for some reason don't want to redo the analysis using custom CDF file.

In this case we select a probeset with highest mean intensity across all the samples in the dataset as a gene representative. I believe it's better than some kind of averaging of different probesets. Any critiques are welcome.

[Will]

not a bad way. I used to try to average the largest subset of "correlated" values. However, I would often end up with two equally sized subsets with opposite values. Then its just a coin-toss as to which one you believe.