[BioC] (no subject)

[Rafael A. Irizarry]

On Mon, 2 Jun 2003 alea at interfree.it wrote:

> 
>  Laurent Gautier <laurent at cbs.dtu.dk> wrote:
> 
> > One should keep in mind the assumption behind many of the normalization
> > techniques: "most of the genes are not differentially expressed across
> > the
> > experiments". Filtering before normalization/scaling should be done with
> > that in mind.
> 
> Hi all.
> I'm a novice,... may be this is he reason why I'm loosing myself..
> What is a "low espressed spot"?
> It seems a problem of logic.
 
a better way to say it is "low intensity spot". low intesities are usually
due to low expression of the gene or probe represented by the spot.

> 
> It seems to me that on one side we say that it not an expressed gene (=
we can not consider as an expressed gene),  it's more likely dued to noise
(background,...).   
> 

but where (and how) do you draw the line? 

> On the other side we say that, however, it's actually an expression and
good for normalizing. 
> Of course a non expressed gene is also not differentially expressed... 

what if its non-expressed on one array and expressed on another? i would
call this differentially expressed.


> 
> Are good enougth for normalization and not enougth for other analysis? 
> Is it another cut-off problem??
> 
> Absent - 
> So and so (for normalizing yes, other analysis no) - 
> Present -
> 

normalization techiniques that use all genes or probes appear to work
succesfully in practice.  papers by dudoit, bolstad, aastrad, li and wong,
huber, and various others demonstrate this. 

> AleA
> 
> 
> 
> -----------------------------------------------------
> 
> Salve, il messaggio che hai ricevuto
> è stato inviato per mezzo del sistema
> di web mail interfree. Se anche tu vuoi 
> una casella di posta free visita il
> sito http://club.interfree.it
> Ti aspettiamo!
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>