Estimation of genetic variance for macro- and micro-environmental sensitivity using double hierarchical generalized linear models

Table 6 The best model selected in 100 replicates according to Akaike’s information criterion and effect of the number of offspring for different true genetic models

NO	True genetic model	True parameters		Nb of times model selected based on AIC
NO	True genetic model	$σ A v 2$	$σ A sl 2$	Macro–micro	Macro	Micro	Simple
100	Macro–micro	0.1	0.05	99	1	0	0
	Macro	0	0.05	5	95	0	0
	Micro	0.1	0	3	1	94	2
	Simple	0	0	1	4	5	90
50	Macro–micro	0.1	0.05	63	27	8	2
	Macro	0	0.05	7	87	1	5
	Micro	0.1	0	4	1	71	24
	Simple	0	0	1	1	3	95

NO = number of offspring per sire, “Macro–micro” = model accounting for both macro- and micro-environmental sensitivities; “Macro” = model with only macro-environmental sensitivity; “Micro” = model with only micro-environmental sensitivity; “Simple” = model without macro- and micro environmental sensitivities and only a genetic effect for the phenotype; $σ A int 2$ = additive genetic variance of breeding value for intercept (true value = 0.3); $σ A sl 2$ = additive genetic variance of breeding value for slope (= macro-environmental sensitivity), $σ A v 2$ = additive genetic variance for environmental variance (= micro-environmental sensitivity), $ρ A int, A v$ = $ρ A int, A sl$ = $ρ A sl, A v$ = 0.