Sense and antisense probes

Question

[abdul77rh]

I am searching public data series in GEO to look into differential gene expression for a particular gene in the context of schizophrenia. For my gene of interest, there are two different probes; one is sense and the other is antisense for the same region of the gene. Expression values for each of the two probes differ widely between available data series, i.e. in certain data series the antisense probe gives much higher expression values, while in others the sense one gives higher expression values.

Looking at the direction of the gene, one would assume that values from the sense probe should be used, however this discrepancy is a problem.

Could anyone help with an explanation please?

[WouterDeCoster]

Your gene of interest could have an antisense transcript, but perhaps it would be helpful to mention which gene you are talking about...

Answer 1

[abdul77rh]

The gene is Neuronal PAS Domain Protein 4 (NPAS4).

[WouterDeCoster]

Please use ADD REPLY to answer to earlier comments, as such this thread remains logically structured and easy to follow. I don't know enough about microarrays to answer your question about having both sense and antisense probes.

[mastal511]

What microarray and platform are you looking at?

There could be several reasons for seeing antisense expression. It's possible that there is true antisense expression from this gene under certain conditions. Another reason could be that the 'antisense' expression may be expression from a gene that is antisense to NPAS4. There could be a problem with the design of the antisense or sense probe sets, and the expression may be due more to the probe sequences than to a true biological signal.

You should try and find out the sequences of the sense and antisense probes, and where along the gene they map.

The reason for the array having been designed with antisense probesets could be, that at the time of array design, there were ESTs in UniGene that mapped to that location on the genome, and it may not have been clear which strand of the genome was being transcribed.

Answer 2

[abdul77rh]

The platform used was Affymetrix Human Genome U133 Plus 2.0 Array The probe sets (both antisense and sense) are correctly labelled. There does not seem to be an antisense transcript of NPAS4, so could it be that the methodology used in the experiment resulted in this apparent discrepancy? Interestingly, I found a similar trend (only antisense probe set giving a significant value) with another data series that was produced in the same year (2008). However, a third data series that was produced a year after (2009) gave a different picture, i.e. the sense probe set gave the significant results! of course, the three data series had similar experimental conditions and were looking at similar type of patients.

[WouterDeCoster]

Again: please use ADD REPLY to answer to earlier comments, as such this thread remains logically structured and easy to follow.

[mastal511]

Are the results given by gene name or by probeset ID? While probe sequences and probeset IDs will not change over time, the probeset annotations may change, especially when there is a new release of the genome assembly.

If you have a look at the Affymetrix web pages, and their NetAffx analysis centre, probeset 1554299_at, the sense probeset, is given an Annotation Grade of A, by Affymetrix, in their current annotations, whereas probeset 240794_at, the antisense probeset, is given Annotation Grade E, meaning that the evidence for the probeset is not as reliable.

I guess the existence of antisense expression for the gene could always be validated by PCR.