Dear members,
I have been ask to evaluate an MA plot where several probes are representing intron/exon spanning regions. This is apparently a sign of genomic dna contamination in the RNA samples. You can find the plot in the link:

https://drive.google.com/file/d/0B82aQne0d_mYMllIYmEyTG9QYmM/view?usp=sharing

What I would like to kindly ask: 
- Is this data reliable?
- Could there be other reasons for the abundance of these probes?
- Is something like this expected in general for microarrays?

Picture is attached.
We are working with the PorGene-1_1-st affimetrix chip. 
We took the annotations from the manufacturer's webisite assigend to this chip.
http://www.affymetrix.com/support/technical/byproduct.affx?product=modelorg_1_1_plate
PorGene-1_1-st Design Time Annotations, Probeset, CSV format (4.4 MB, 10/31/11)

Thank you

It is much more likely that what you are seeing are unprocessed mRNA sequences, rather than genomic DNA. There are also a set of intronic negative controls on Gene ST arrays, and these have the most unfortunate habit of popping up in the set of 'top hits' if you don't first filter them out. This is a problem for both the Gene ST arrays, and RNA-Seq if you don't first do poly-A selection (in my experience, up to 25-30% of the reads in a non-poly-A selected RNA-Seq experiment will map to intronic regions).

In other words, when mRNA is first transcribed, it still contains all the intronic sequences, which are then excised. After that step the poly-A tails are ligated, and the mRNA is sent out to be translated to protein. At any given moment there are any number of these unprocessed mRNA transcripts floating around, and if you purify RNA, and then use random primers to reverse-translate back to cDNA, then you will inevitably generate cDNA that consists of exon-intron boundary sequences, as well as pure intronic sequences. But these come from mRNA, not DNA. If there were DNA contamination (in the case of RNA-Seq), you would get lots of inter-genic regions as well, but that is in my experience really rare.

The old school 3'-biased Affy arrays didn't have this problem because they used oligo-dT as the primer for the reverse transcription step, which is in essence a poly-A selection procedure.

Thank you very much James,

I have spoken with the technician that performed the experiment. He told me two things:

- They did oligo-dT reverse transcription

- The did not do DNAse digestion.

Could it be that among the oligo-dT transcribed cDNAs,  ployA-inron-exon mRNAs are present?

Could lack of DNAse treatment explain the intron/exon probes presence?

Thanks