[BioC] retrieve expression of all probes in the probe set

James W. MacDonald jmacdon at uw.edu
Mon Oct 15 17:36:47 CEST 2012

Hi Andrea,

On 10/15/2012 11:01 AM, andrea.grilli at ior.it wrote:
> Dear BioC list,
> I'm performing an analysis with Gene Chip hgu133a plus 2.0 array in 
> two different cell lines. I'm interested in the expression of a 
> specific gene and two different probesets match it: despite one is 
> saying no difference exists, the other is suggesting a strong 
> down-regulation.
> Checking where the two probes match, the first is mapping on the 3' 
> UTR region of the transcript, instead the second has each one of the 
> probes composing the probe set on different exons.
> So two questions:
> - Could be one probe is more reliable than the other? which one and why?

First, I am assuming that when you say probe you really mean probeset. 
These are different things.

The 3'biased probeset is probably more reliable. The IVT step for these 
arrays uses oligo-dT as a primer, thus you are starting at the poly-A 
tail. So two things; you are starting the IVT at the poly-A tail, so the 
likelihood of creating cDNA goes down the farther 5' you go (e.g., early 
termination will result in more cDNA near the 3' end and less near the 
5' end).

Since the IVT uses the poly-A tail, by definition you are using mRNA 
that if degraded, will only be degraded on the 5' end. You can look at 
the RNA degradation plots to show that the farther 5' you go, the lower 
the expression levels.

> - Do yuo know if it is possible to recover the expression of each 
> probe of the probe set? Different splicing variants for this gene 
> exist, so I want to check if the expression of each single exon can 
> tell me more.

Yes, you can get the expression values from your AffyBatch. One could 
argue that any data should be first background corrected and normalized 
before extracting, so you might first do:

bg.abatch <- bg.correct(abatch, "rma")
bg.norm.abatch <- normalize(bg.abatch, "quantiles")

and then you can extract the PM probes using the aptly-named pm() function.

pm(bg.norm.abatch, <probeset ID goes here>)



> Thanks in advance for your help,
> Andrea
> _______________________________________________
> Bioconductor mailing list
> 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.
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099

More information about the Bioconductor mailing list