Both that discussion and this one are related to my questions. Phylogenetic models are analogous to pedigree modes, and upon further reading I think the best way to address my second question will be to use the hypothesis framework from the phylogenetic model vignette.
I am still trying to figure out whether to use the identity phi to calculate residual variance (π2π) when the genetic variance (π2π) is estimated on the logit link scale.