[BioC] questions about MVA plot and loess N

[Rafael A. Irizarry]

On Wed, 25 Jun 2003 Phguardiol at aol.com wrote:

> Hi,
> few questions:
> 1- could it be possible to get the IQR values separately when using the
mva.pairs function ? That would be useful if a large number of chips are
run at the same time and one wants to compare the different normalization
methods ?   

you can use, for example, a for loop and the function IQR

> 2- is there a way to modify the name of the chip inside the MA plot ?

read the man page.

> 3- should I care about getting with loess normalization the following
warning message: k-d tree limited by memory . ncmax=5002 9exact process
was: data2 <- expresso(data, normalize.method="loess",
bgcorrect.method="rma", summary.method="medianpolish",
pmcorrect.method="pmonly")     

no. just warnings from the function loess.


> 4- where can I find more informaiton about the different summary.method
available in affy ? some are completely unkown to me and I d like to learn
about these   

the help files and vignette have some references.

> 5- about normalization what I have understood is that some tools like MA
plots can tell you that you have not "normalized enough" your data, but is
there a way to see that you have gone too far with the process ? Maybe
that s a stupid question....sorry !   > thanks for your help anyway
> Philippe 

not stupid question at all. its hard when you dont have some genes for
which you now there should be a difference. for more on this read the
2003 bioinformatics paper by bolstad et al. 

> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>