[BioC] limma and normexp

[Gordon K Smyth]

> Date: Tue, 11 Jan 2005 12:52:15 -0000
> From: "Helen Cattan" <helen.cattan at jenner.ac.uk>
> Subject: [BioC] limma and normexp
> To: <bioconductor at stat.math.ethz.ch>
>
>
> Hi,
>
> R v 2.0.1 and Limma v 1.8.15
>
> Can anyone please explain to me why I get this error when trying to use
> the background correct method normexp. I have no missing values and as
> far as I can tell from looking at summary information this array is not
> very different from all the others. Other backgroundCorrect methods work
> fine.
> Code with error is below.
> Thanks,
>
> Helen
>
> library(limma)
> targets=readTargets()
> myfun=function(x) as.numeric(x$Flags>1)
> RG=read.maimages(targets$FileName, source="genepix", wt.fun=myfun)
> RG$printer=getLayout(RG$genes)
> RG$genes=readGAL("M15k4.gal")
> spottypes=readSpotTypes()
> RG$genes$Status=controlStatus(spottypes, RG)
> RGnormexp=backgroundCorrect(RG, method="normexp", offset=50)
> Corrected array 1
> Corrected array 2
> Corrected array 3
> Error in fit.normexp(foreground = RG$G[, j], background = RG$Gb[, j]) :
>         optim failure: degeneracy of the Nelder-Mead simplex

This is a failure in the iterative algorithm used to estimate the background correction.  The
other background correction methods are not iterative.  Sorry I don't have a solution at this
time.  We'll continue to work on it to make the estimation algorithm more bullet-proof.

Gordon