[R-sig-eco] Lynch 1991 MCMCglmm?
mailbox Franz
f.krah at mailbox.org
Tue Feb 21 13:29:22 CET 2017
Hello list,
I want to run the Lynch1991 method in MCMCglmm.
Due to the size of the dataset (2500 species), MCMCglmm was suggested to me.
(ape::compar.lynch does not provide likelihoods, after several tries with different EPS.)
I want to run trait partitioning into a phylogenetic and a specific effect for a single trait.
However, I am not sure how to specify that in MCMCglmm.
This model specification seems helpful, however, I am not sure if this is the correct way.
From the output I am no able to derive the estimates I want:
- grand mean
- deviation from the grand mean
these together are the phylogenetic effect
and
- residuals as the specific effect.
MCMCglmm(phenotype~1, random=~taxon, pedigree=list(taxon=Ainv),
data=test.data, prior=prior, verbose=FALSE)
Hope someone can help or share experiences!
Cheers,
Franz
More information about the R-sig-ecology
mailing list