[R-sig-ME] Is there any way to model nested random effects in MCMCglmm?

[Jarrod Hadfield]

Hi,

Stating that terms are explicitly nested is not necessary as long as  
everything is given a unique identifier. For example, you don't have  
the same trait within a species. If you do, you can just relabel it.  
Then use:

random~Species+Population+Trait

Cheers,

Jarrod


Quoting Ivain MARTINOSSI--ALLIBERT  
<ivain.martinossi--allibert at agroparistech.fr> on Fri, 29 Mar 2013  
16:20:50 +0100 (CET):

> Hello,
> I have data of quantitative gentics composed of various traits  
> measurement for multiple population and multiple species.
> In each species one can find few populations and in each population  
> a few traits are measured, so I initially constructed a glmer model  
> as follows:
>
> model<-glmer(response~fixed effetcs+ (1|Species/Population/Trait) , data=a)
>
> I needed then to add phylogeny to my analysis and decided to use  
> MCMCglmm. I realised I was unable to take into account the nested  
> structure of the three random effects:
> Species, Population and Trait
> If anyone has an idea of how it should be done, I would be very  
> grateful for help.
>
> 	[[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.