[BioC] MAANOVA comparison matrix question

Cecilia McGregor cmcgre1 at lsu.edu
Tue May 6 16:49:04 CEST 2008


Hi All

I have 2 cultivars (bx and nas), 2 treatments (s and h), 4 timepoints (T1 to T4), and 4 biological reps for each cultivarxteatment. My experiment consists of 32 two-color arrays hybridized in 8 loops. Every EST is spotted in duplicate.

For starters I want to compare just within one treatment, so just for bx - I want to compare treatments over time. Then I want to do the same for the other treatment (nas). Once I have this I want to compare within treatments. So first just for treatment s, between nas and bx. 

So far I've done the following: *don't worry about the low n.perm. I'm just testing things at this stage and will use more permutations once I think I know exactly what I need to do).

#Load data
bnhsnoflags1.raw<-read.madata("bnhsnoflags1.txt",designfile="design.txt",probeid=1,metarow=2,metacol=3,row=4,col=5,intensity=6,arrayType='twoColor',log.trans=TRUE,spotflag=FALSE,n.rep=2)
 
#transform using rlowess normilization
bnhsnoflags1<-transform.madata(bnhsnoflags1.raw,method="rlowess")

#close graphics
graphics.off()

# Fit Full ANOVA mixed model
anova.full.mix<-fitmaanova(madata=bnhsnoflags1,formula=~Array+Dye+Array:Dye+Cultivar+Treatment+Time+Cultivar:Treatment+Cultivar:Time+Treatment:Time+Sample+Spot,random=~Array+Sample+Spot)
 
>From what I saw on p.21 of the April 25 2007 manual I think I should now do the following

#Test Statistic (F and Fs by default)
ftest.int.mix<-matest(data=bnhsnoflags1,anovaobj=anova.full.mix,term="Treatment:Time",n.perm=5)

and then make a comparison matrix. Is the treatment:time interaction the correct one to test differences between treatments for bx over time?

Any tips about making a comparison matrix?

Then after the comparison matrix  I do a ttest (p.22 of manual) on this term? And then I can change the comparison matrix to test the same for nas?

Is this correct? I guess the first question is: is Treatment:Time the correct interaction to test for what I want? 
And where can I find more help/info about making the comparison matrix?

Thanks

Cecilia





Cecilia McGregor, Ph.D.
Sweetpotato Breeding and Genetics Lab
School of Plant, Environmental and Soil Sciences
236 J.C. Miller Hall
Louisiana State University
Baton Rouge
LA, 70803
USA

Phone: (225) 763-9138
Fax: (225) 578-1068



More information about the Bioconductor mailing list