Skip to content
Prev 56 / 20628 Next

incorporate pedigree into lmer

No,

The animal model requires specification of the variance-covariance 
matrix of the random effects, according to the pedigree (kinship) 
information. lmer and lmer2 currently do not allow this, as far as I 
know. (There is the pedigree constructor function in package lme4, as 
you noticed, so presumably this functionality is on the way.) You could 
perhaps use lme (package nlme) to fit the model, but specifying the 
var-covar matrix for such complex data will be painful. There is lmekin 
in the kinship package, which may be more useful to you. However, I 
haven't used it so I can't be of further help. I don't know how to 
estimate separate additive and dominance matrices. Maybe someone more 
experienced than me will reply.

I think there is a need for a quantitative genetics package for R, 
although some qtl mapping methods seem to be implemented (See the 
Genetics task view on CRAN). So much statistics, so little time. *sigh*

Simon.
There is Doran, Harold wrote: