Table 4.

Proportion of simulated data assigned to each of the four possible trio relationships, at a threshold of 0.9 for Nucl3ar. In the first scenario, 1000 trios were analyzed of which 200 and 800 trios are simulated from two populations with levels of population differentiation similar to the HapMap Japanese and Chinese respectively, which we have loosely defined to reflect fine-scale population differentiation; in the second scenario, 1000 trios were analyzed of which 200 and 800 trios are simulated from two populations with levels of population differentiation similar to the HapMap CEPHs and Yorubas respectively, which we have loosely defined to reflect broad-scale population differentiation. There are no detectable differences in the performance of Nucl3ar in both scenarios.

	Assigned Relationship
	True	Misspecified parents	Misspecified-father	Misspecified-mother
Fine-scale
Pop. 1 (200 trios)	0.972	0.000	0.021	0.007
Pop. 2 (800 trios)	0.963	0.000	0.024	0.014
Broad-scale
Pop. 1 (200 trios)	0.987	0.000	0.013	0.000
Pop. 2 (800 trios)	0.992	0.000	0.002	0.006