[BioC] RNA degradation plot of transcript arrays

[James W. MacDonald]

Hi Assa,

On 10/30/2012 5:18 AM, Assa Yeroslaviz wrote:
> Hi everyone,
>
> We are working on the affymetrix transcript arrays (HuGene-1_0-st-v1). As a
> first step I ran a quality control check. One of the steps was the RNA
> degradation plot.
>
> As I know it from the other Oligonucleotide arrays, in the best case
> scenario the mean intensity is low at the 5' end and gets higher towards
> the 3' end.
> What I got in this analysis was a different picture (see attachment).
>
> Here it goes higher and in the middle down again. Is this behaviour a
> normal situation for transcript arrays?

Yes. The 3'-biased arrays all used oligo-dT as primer, so the IVT step 
always started at the 3' end of the mRNA. This gave a 3' bias for two 
main reasons; first, you start the IVT at the 3' end, so if the 
transcription fails (on average) before reaching the 5' end, you will 
have comparatively more sequence from the 3' end.

Second, by definition any mRNA that is transcribed using oligo-dT isn't 
degraded at the 3' end (because if it were, the primer wouldn't attach). 
But there is no restriction on degradation from the 5' end. So the mRNA 
species you are transcribing are biased in that they are all not 
degraded at the 3' end, but may well be degraded to a greater or lesser 
extent on the 5' end.

However, with the ST arrays, we are now using random primers, so we are 
no longer biasing towards either end. In fact we are biasing AGAINST the 
ends, because we can use mRNA that is being degraded (on average) from 
either end. If the transcripts are not degraded at all, (and we assume 
the IVT step transcribes from the primer to the end of the transcript on 
average), then we should expect the RNA degradation plot to be exactly 
opposite of what you see with the 3' biased arrays.

Best,

Jim
>
> Can someone point me to a paper or an explanation for this behaviour?
>
> Thanks a lot
>
>
> Assa
>
>
> _______________________________________________
> 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.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099