[BioC] [MAANOVA] Warnings in fitmaanova & information request

[Loren Engrav]

Your query is beyond me and the maanova documentation is a little skimpy for
"complex" models. The maanova people have a tutorial that provides much more
info which I found extremely useful. I emailed them asking if they wish to
share the tutorial. And, in keeping with their previous responses, they
responded even before I sent this. So check out the discussion R/Maanova
Tutorial.


> From: Maciej Jończyk <mjonczyk at biol.uw.edu.pl>
> Reply-To: Maciej Jończyk <mjonczyk at biol.uw.edu.pl>
> Date: Sat, 13 Mar 2010 17:54:39 +0100
> To: Loren Engrav <engrav at u.washington.edu>
> Cc: Maciej Jończyk <mjonczyk at biol.uw.edu.pl>, "bioconductor at stat.math.ethz.ch"
> <bioconductor at stat.math.ethz.ch>
> Subject: Re: [MAANOVA] Warnings in fitmaanova & information request
> 
> Thank you Loren, I'll ignore this error.
> 
>> From: Loren Engrav <engrav <at> u.washington.edu>
>> Subject: Re: [MAANOVA] Warnings in fitmaanova & information request
>> Newsgroups: gmane.science.biology.informatics.conductor
>> Date: 2010-03-12 02:07:33 GMT (1 day, 14 hours and 34 minutes ago)
>> 
>> Greetings
>> 
>> If you are asking is this warning significant
>> 
>>> Warning messages:
>> 1: In any(result$random[idx.mainterm]) :
>>  coercing
>>> argument of type 'double' to logical
>> 2
>> 
>> I have used maanova many many times and see that message every time and
> the
>> maanova people indicated I should ignore it
> 
> But I have next problem and I want to ask for advice.
> 
> *
> I pasted some of the previous supplied information (I can't respond from
> list level,
> I respond to verification message but it was a day ago and nothing
> happens :( ).
> 
> 
> #####################################################################
> I'm trying to analyse simple loop-design in MAANOVA package.
> I've used two-colour spotted oligo arrays.
> 
> I have two maize lines (linia): dl, dh;
> two conditions (temp): c (cold), k (control).
> 
> Here is my design:
> 
> dlc ---- dlk
> |    X    |
> dhc ---- dhk
> 
> with two more hybridizations, dlc vs dhk and dlk vs dhc ("X" in this
> scheme).
> 
> I repeated this scheme four times, with dye swap (two labelings in
> both directions).
> 
> Technical replication is within each replicated scheme. So within
> scheme all instances of each sample (samples: dlc, dlk, dhc, dhk)
> is from the same isolation (i.e. in each loop each RNA is used
> three times in different hybridizations).
> 
> Here is my design file (powt_b is a number of biological replicate):
> 
> Array  Dye Sample linia temp powt_b
> dhk_dhc093 Cy5 1 dh c 1
> dhk_dhc093 Cy3 2 dh k 1
> dhc_dhk104 Cy5 3 dh k 2
> dhc_dhk104 Cy3 4 dh c 2
> dhk_dhc116 Cy5 5 dh c 3
> dhk_dhc116 Cy3 6 dh k 3
> dhc_dhk016 Cy5 7 dh k 4
> dhc_dhk016 Cy3 8 dh c 4
> dhk_dlk094 Cy5 9 dl k 1
> dhk_dlk094 Cy3 2 dh k 1
> dlk_dhk105 Cy5 3 dh k 2
> dlk_dhk105 Cy3 10 dl k 2
> dhk_dlk117 Cy5 11 dl k 3
> dhk_dlk117 Cy3 6 dh k 3
> dlk_dhk139 Cy5 7 dh k 4
> dlk_dhk139 Cy3 12 dl k 4
> dlc_dlk092 Cy5 9 dl k 1
> dlc_dlk092 Cy3 13 dl c 1
> dlk_dlc106 Cy5 14 dl c 2
> dlk_dlc106 Cy3 10 dl k 2
> dlc_dlk118 Cy5 11 dl k 3
> dlc_dlk118 Cy3 15 dl c 3
> dlk_dlc023 Cy5 16 dl c 4
> dlk_dlc023 Cy3 12 dl k 4
> dlc_dhc095 Cy5 1 dh c 1
> dlc_dhc095 Cy3 13 dl c 1
> dhc_dlc107 Cy5 14 dl c 2
> dhc_dlc107 Cy3 4 dh c 2
> dlc_dhc119 Cy5 5 dh c 3
> dlc_dhc119 Cy3 15 dl c 3
> dhc_dlc136 Cy5 16 dl c 4
> dhc_dlc136 Cy3 8 dh c 4
> dlk_dhc101 Cy5 1 dh c 1
> dlk_dhc101 Cy3 9 dl k 1
> dhc_dlk103 Cy5 10 dl k 2
> dhc_dlk103 Cy3 4 dh c 2
> dlk_dhc121 Cy5 5 dh c 3
> dlk_dhc121 Cy3 11 dl k 3
> dhc_dlk015 Cy5 12 dl k 4
> dhc_dlk015 Cy3 8 dh c 4
> dhk_dlc100 Cy5 13 dl c 1
> dhk_dlc100 Cy3 2 dh k 1
> dlc_dhk102 Cy5 3 dh k 2
> dlc_dhk102 Cy3 14 dl c 2
> dhk_dlc120 Cy5 15 dl c 3
> dhk_dlc120 Cy3 6 dh k 3
> dlc_dhk140 Cy5 7 dh k 4
> dlc_dhk140 Cy3 16 dl c 4
> 
> As required, Sample denotes distinct RNAs. So, e.g. condition
> "dl c" has the same number within single replication of entire
> scheme (what makes three technical replicates of RNA withnin
> biological replicate).
> 
> I wrote the model. I want to include factors linia, powt and
> their interaction as a fixed effects. I also included Sample,
> powt_b and Array as a random effects. In this way I want to
> include effect of techical replication. I didn't include Dye
> effect, because the experiment was dye balanced, and data
> normalized.
> 
> ######################################################################
> 
> After "fitmaanova":
> 
> 
>> model.full_nodye=fitmaanova(eth_logrg,formula=~Array+Sample+powt_b+linia+temp
> +linia:temp,random=~Array+Sample+powt_b)
> 
> 
> I wanted to set up contrasts for interaction terms:
> 
> dlc vs dlk
> dhc vs dhk
> dlc vs dhc
> dlk vs dhk
> 
> I constructed design matrix (I suppose that variants are ordered
> alphabetically).
> 
>> cont_matx=matrix(c(1,-1,0,0,0,0,1,-1,1,0,-1,0,0,1,0,-1),nrow=4,byrow=T)
> 
> And I tried to perform test
> 
> 
>> test=matest(eth_logrg,model.full_nodye,term="linia:temp",Contrast=cont_matx,
> n.perm=61,shuffle.method="resid")
> 
> But it gave error message:
> 
> Error: The number 1 test is not estimable
> 
> What is wrong? Am I included too many effects in fitmaanova?
> I have four replications of the experiment, I think it is enought.
> Or maybe my contrasts' specification is wrong?
> I haven't a clue what is wrong.
> 
> Help will be very appreciated.
> 
> Best wishes,
> 
> Maciej
> 
> 
> Maciej Jończyk PhD
> Department of Plant Molecular Ecophysiology
> Institute of Plant Experimental Biology
> Faculty of Biology, University of Warsaw
> 02-096 Warszawa, Miecznikowa 1
> 
> 
> 
> ___________________________________
> NOCC, http://nocc.sourceforge.net
> 
> 
> 
> 
> -- 
> This email was Anti Virus checked by Astaro Security Gateway.
> http://www.astaro.com