Table 2.

Results of the D_4P test. The results are from NA19239 (for YRI), NA12891 (for CEU), HG00421 (for ASN), and the Aboriginal Australian genome (ABR). The two groups are patterns representing the two ways in which eligible SNPs can partition the four genomes (they have not been polarized).

	Group 1	Group 2
YRI	1	1
ABR	0	1
CEU	0	0
ASN	1	0
Observed number*	13,974	14,765
Observed proportion (95% CI)†	48.6% (47.8 to 49.4%)	51.4% (50.6 to 52.2%)
Expected proportion under multiple-dispersal model 1‡	48.7%	51.3%
Expected proportion under multiple-dispersal model 2§	48.0%	52.0%
Expected proportion under single-dispersal model‖	50.3%	49.7%

^*Average number of eligible SNPs showing groups 1 and 2 across block bootstrap replicates.

^†95% confidence interval obtained from a block bootstrap (4). Z test rejects the null hypothesis that this value is equal to 50% (Z = 3.3, P < 0.001).

^‡Expected proportion from a multiple-dispersal model in which aboriginal Australasians split from Eurasian populations 2500 generations ago, before the split of European and Asian populations. This split time was estimated using the Aboriginal, NA12891, and HG00421 sequences (4). These were the same individuals used for the D_4P analysis.

^§Expected proportion from a multiple-dispersal model in which aboriginal Australasians split from Eurasian populations 2750 generations ago, before the split of European and Asian populations. This split time was estimated using the Aboriginal Australian and all Eurasian sequences (4).

^‖Expected proportion from coalescent simulations under a model in which aboriginal Australasians split from Asian populations 1500 generations ago. The other parameters were those estimated by Schaffner et al. (27). See (4) for additional models.