[BioC] S/N ratio in microarrays

[Naomi Altman]

I have not reviewed the literature either, but I have analyzed 
experiments on 2-color cDNA, 2-color oligo Agilent, Affy and 
Nimblegen arrays, both model species and custom arrays for non-model 
species, with a variety of experimental designs.

Based on this experience, I would have to agree with Sean.  Platform, 
species, handling of the biological material, experience of the lab 
personnel, micro-environment, ... make a huge difference.  But the 
S/N is certainly orders of magnitude smaller than 30000:1 - maybe 
more like 2:1 on a single array.  Of course, partly that is because 
of the definition of signal - do you mean the level of mRNA in the 
sample (which I think you mean in spectroscopy), or the mean level 
under the conditions of interest (which is what we mean in most 
microarray studies)?

--Naomi

At 07:04 AM 11/10/2005, Sean Davis wrote:
>On 11/10/05 6:18 AM, "narinder.singh at diagenic.com"
><narinder.singh at diagenic.com> wrote:
>
> > Can somebody point to litterature related to S/N ratio in 
> microarray data. In
> > spectroscopy the S/N ratio is typically 30000:1. From what I have 
> gathered S/N
> > ratio is a big problem in microarray data. Is there any 
> litterature/study with
> > corresponding S/N numbers for microarray data for commercially available
> > platforms.
> >
> > Thanks in advance.
>
>Narinder,
>
>I don't think this is a well-understood property of array technology.  S/N
>is at least probe-dependent and like sample-dependent as well (RNA quality,
>etc.), not to mention operator dependent, ozone-level dependent, humidity
>dependent, protocol dependent, and many others.  Some experimental designs
>(particularly with two-color arrays) can GREATLY affect the S/N ratio at the
>level of the final analysis (which is what counts).  I'm not giving an
>answer directly, because I'm not sure that one exists and if it does, I
>think it applies only to a specific set of samples in a specific lab on a
>couple of platforms, etc.
>
>Perhaps someone else has reviewed the literature on this subject and can
>give you a better answer.
>
>Why do you ask?
>
>Sean
>
>_______________________________________________
>Bioconductor mailing list
>Bioconductor at stat.math.ethz.ch
>https://stat.ethz.ch/mailman/listinfo/bioconductor

Naomi S. Altman                                814-865-3791 (voice)
Associate Professor
Dept. of Statistics                              814-863-7114 (fax)
Penn State University                         814-865-1348 (Statistics)
University Park, PA 16802-2111