[R-sig-ME] MCMCglmm prior specification

[Jarrod Hadfield]

Hi Tanya,

The error message is because your between-line covariance matrix is  
6x6 (each sex at three temperatures) but your prior is for a 2x2  
covariance matrix. With 14 lines I would not advise fitting a 6x6  
covariance matrix (21 parameters) and I would try and keep it simple.

Cheers,

Jarrod





Quoting Tanya Pennell <T.Pennell at sussex.ac.uk> on Wed, 23 Jul 2014  
10:16:50 +0000:

> Hi all,
>
> I'm trying to amend a previous model that included gene x sex  
> interactions, but now i'm looking at gene x sex x environment  
> interactions. I'm getting an error message appear that says 'V is  
> the wrong dimension for some prior$G/prior$R elements'
>
> Here is my code -
>
>
> prior.model.2<-list(R=list(V=diag(2)/252, nu=0.01),G=list     (   
> G1=list   (V=diag(2)/252, nu=2, alpha.mu=c(0,0),  
> alpha.V=diag(2)/252)))
>
>
> model.2 <- MCMCglmm(relative_fec ~  
> sex*rep-1,random=~us(sex:temp):line, rcov=~idh(sex):units,  
> family="gaussian", nitt = 100000, burnin = 30000, thin=30, data =  
> all_treatments, prior = prior.model.2, verbose = FALSE)
>
>
> I have relative fecundity measurements for males and females from 14  
> lines, that were measured in 3 temperatures. Each line, sex and temp  
> combination had  6 replicates.
>
> Can anyone guide me on my prior here please?
>
> Many thanks,
> Tanya
>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>


-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.