[BioC] maanova: missing observations in cov/cor

Matthew Vaughn vaughn at cshl.edu
Mon Aug 21 21:16:30 CEST 2006


I am trying to get a handle on using maanova. I've managed to get everything
sorted up to the point of running matest. I've seen the particular problem I
will describe posted here before and there's never been a satisfactory solution
that I can see.

model.mixed.2 <- makeModel(data=myData.norm,
formula=~Array+Dye+Treatment+Sample +Array:Sample+Array:Treatment,
random=~Array)

anova.mix.2<-fitmaanova(myData.norm, model.mixed.2)

test.Treatment<-matest(myData.glowess,
model=model.mixed.2,term="Treatment",n.perm=100)

dies with:

Error in var(log(Chis)) : missing observations in cov/cor In addition: Warning
message: NAs produced in: rchisq(n, df)

I've gone over my data and I don't see any missing values. I had a look at the
primary maanova source code and don't see where this error is generated, so I am
a bit stuck.

Any advice?



More information about the Bioconductor mailing list