[BioC] Exclude probes that show sd above 0.1 between replicatevalues

[J.Oosting at lumc.nl]

Hi João

Here's how I did it on a set of arrays with 3 replicates. ID is a column that has identical values for replicate spots. 

spotaverage<-function(x) {
  # at max 1 NA, and have low variability
  if (sum(is.na(x))<2 && sd(x,na.rm=TRUE)<0.1)  median(x,na.rm=TRUE)
  else NA
}
aggcollumn<-function(x) {
  agg<-aggregate(x,list(genes=MAn$genes[,"ID"]),FUN=spotaverage)
  ac<-as.numeric(agg[,2])
  names(ac)<-agg[,1]
  ac
}

avg.m<-apply(MAn$M,2,aggcollumn)

 
> 
> Dear list, 
>  
> I have a MAList and I want to exclude probes that show a 
> standard deviation above 0.1 between the replicate values. 
> The number of within-array replicates for each probe (ndups) 
> is equal to 2 and the number of spots to step from a probe to 
> its duplicate (spacing) is equal to 1.
> I am not able to to this. Can somebody give me some a hint to 
> resolve this?
>  
>  
> Best regards
> 
> João Fadista
> Ph.d. student
>