[BioC] question of medianpolish in RMA

[Ben Bolstad]

There is no such constraint on signs in the algorithm itself. However,
remember that the expression values returned from RMA are in log2 scale
so negatives are not really negatives.

That said, it is not common in my experience to ever see negative values
(even in the log2 scale) being returned. If such a thing is happening
for your data it might be indicative of some other problem.




On Mon, 2005-05-30 at 16:34 -0500, Simon Lin wrote:
> How can the medianpolish used in RMA ensure that
>  both
>    alpha (probe affinity) and beta (expression abundance)
>  larger than zero?
> 
> Or, RMA occasionally gives out negative expression values?
> 
> Thanks!
> 
> Simon
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
-- 
Ben Bolstad <bolstad at stat.berkeley.edu>
http://www.stat.berkeley.edu/~bolstad