Least-squares additive genetic (co)variance of the current population: the
covariance matrix of the best linear prediction of genotypic values from
the state of every locus (the sum of its two alleles: their effect vectors,
or for binary_b() their effects B y, or for contribution_matrix() the
gene values). For the architectures in evomod2 the genotypic value is
linear in the state of each locus given the others, so this is the
additive variance in the usual sense (variance of breeding values) under
linkage equilibrium. G - additive is the non-additive (epistatic)
covariance. Needs more individuals than loci times traits.
Value
A list with G, additive and nonadditive (G - additive), all
with the n - 1 denominator of stats::cov().
Examples
eps <- pairwise_epsilon(loci = 10, traits = 2, sd = 1)
m <- evomod_model(multilinear(10, 2, eps), mutation(0.01, 0.2), env_sd = 1, init_sd = 0.1)
pop <- evolve(m, stabilizing(200, omega = 20), n = 300, seed = 1)
additive_covariance(pop)
#> $G
#> trait1 trait2
#> trait1 1.8910878 0.1247169
#> trait2 0.1247169 1.6747977
#>
#> $additive
#> trait1 trait2
#> trait1 0.9021293 0.1499882
#> trait2 0.1499882 0.8469063
#>
#> $nonadditive
#> trait1 trait2
#> trait1 0.98895850 -0.02527132
#> trait2 -0.02527132 0.82789140
#>