[BioC] error in mas5calls() for soybean chip

[Ken Simpson]

Hi Jenny,

>  > PMAcalls <- mas5calls(rawdata[,1:5])
> Getting probe level data...
> Computing p-values
> Making P/M/A Calls
> Error in if (y < alpha1) { : missing value where TRUE/FALSE needed
>
>  > PMAcalls <- mas5calls(rawdata[,1:5],alpha1=0.04,alpha2=0.06)
> Getting probe level data...
> Computing p-values
> Making P/M/A Calls
> Error in if (y < alpha1) { : missing value where TRUE/FALSE needed

I've seen this error with soybean chips myself, and traced it to probe sets
which contained only a single probe (might also have occurred for those with 2
probes).  I think the Affy P/A call fails for such cases, and is probably
returning NAs.  You could try removing these probe sets before calling
mas5calls(), or writing an alternative version of mas5calls() that handles the
NAs more gracefully.

Cheers,

Ken