TABLE 4.

Comparison between simulation results and predictions for the MSD additive variance after bottlenecks

Model		k	N = 1000	N = 50	N = 10	N = 2
I	Simulation	4	0.0019	0.0041	0.0060	0.0045
		2	0.0019	0.0041	0.0059	0.0039
		1	0.0018	0.0041	0.0059	0.0036
	Diffusion	1	0.0018	0.0036	0.0058	0.0029
	Shortcut	1	0.0019	0.0035	0.0045	0.0021
II	Simulation	4	0.0024	0.0049	0.0067	0.0048
		2	0.0024	0.0045	0.0058	0.0038
		1	0.0024	0.0045	0.0056	0.0035
	Diffusion	1	0.0023	0.0040	0.0050	0.0025
	Shortcut	1	0.0024	0.0039	0.0043	0.0020
III	Simulation	4	0.0079	0.0129	0.0128	0.0066
		2	0.0079	0.0122	0.0115	0.0048
		1	0.0080	0.0119	0.0109	0.0044
	Diffusion	1	0.0076	0.0103	0.0099	0.0034
	Shortcut	1	0.0080	0.0100	0.0085	0.0023

Values of the MSD additive variance obtained by simulation in the base population (N = 1000) or after 100 bottlenecks (N = 2, 10, 50, estimated from samples of 100 offspring in each line), for the different models and types of gene action (k = 1, no epistasis; k = 2, moderate synergistic epistasis; k = 4, strong synergistic epistasis), and MSD expectations (k = 1) derived from diffusion theory or a shortcut method (García-Dorado 2007) are shown.