[BioC] Does the strand of a microarray probe matter?

Charles Danko dankoc at gmail.com
Thu Nov 20 19:10:46 CET 2008


Hi,

>Hi Cei,
>
>The Illumina BeadArrays use a variety of assays.
>
>The whole genome gene expression arrays such as the HumanWG-6 and
>HumanRef-8 use a 3' based
>assay (IVT or in-vitro transcription assay) very similar to the one used
>with the older 3' Affymetrix arrays.
>
>Some of the older or custom gene expression arrays which use Illumina's
>universal BeadChips use the DASL assay, which uses a combination of
>poly(T) priming and random priming.
>
>
>However, if we are talking about gene expression arrays, both types of
>protocol start with an RNA sample, so they will only amplify transcripts
>that are expressed, whichever strand they are transcribed from.
>
>
>The problem with probes on the wrong strand is that you will not be able
>to detect expression from the gene in question if the probe is on the
>opposite strand. If a probe that is on the wrong strand gives a signal
>that indicates expression, then what is really being expressed is possibly
>an antisense transcript which may have some regulatory function.

At least for the Affymetrix arrays, RT-PCR confirms that many of the
probes anti-sense to the gene of interest target anti-sense RNA
transcripts (See this article:
http://www.biomedcentral.com/1471-2164/8/200).  Even in the (slightly)
newer human U133A/2.0 arrays, many probes are targeted to the wrong
strand.  Given the wide use of these Affymetrix arrays, and the wide
availability of data in GEO and other repositories, this is
potentially a HUGE resource waiting for someone with a little bit of
extra time.

>
>Best wishes,
>Maria



More information about the Bioconductor mailing list