Selecting Probes When Rna Degradation Is Obvious

Question

[Julien Textoris]

Hi everyone,

I am analyzing a dataset where there is obvious 5' end RNA degradation. It was known before hybridizing the chip and we could not avoid this. In affy package, there is a function AffyRNAdeg() which helps you assessing this trend.

But now that i'm sure that i have 5' end degradation, what can I do ? I was wondering if in the summarization process I could use only the 4 or 5 PM probes located at the 3' end, for example, but I did not find a function already dealing with that. Is anyone aware of such a way of summarizing affy data ? or do I have to write it myself ?

Thanks

Julien

Answer 1

[Sean Davis]

Unless the degradation is non-uniform (affects only a portion of the samples), I think the general approach is to normalize as usual. The slope of the RNA degradation plot is really not that important (personal communication), only the relative slopes of all the samples in the dataset. Only you can make the decision, but I'd suggest trying normal processing of the samples first before trying to do something exotic.

[Julien Textoris]

thanks for this ! I'll do the usual way then.

Julien