Skip to main content

Table 4 Results from admixture simulation with changing parameter in the HI model from HapMap III data

From: Detecting individual ancestry in the human genome

Parameter

sNMF (R2)

ADMIXTURE (R2)

fastSTRUCTURE (R2)

Sample size

   

8

98.2

99

86.66

10

99.5

99.52

98.69

20

99.74

99.82

99.71

40

99.85

99.9

99.86

50

99.87

99.93

99.9

100

99.91

99.95

99.95

nsnps

   

5

4.56

15.38

19.44

10

15.92

47.37

46.2

50

80.62

86.31

86.89

100

89.67

93.04

93.33

500

98.46

99.07

99.11

1,000

99.19

99.54

99.56

5,000

99.84

99.92

99.91

10,000

99.91

99.95

99.95

Nbreaks

   

5

88.82

88.37

87.46

10

94.38

94.86

94.43

50

98.74

98.87

98.8

100

99.33

99.41

99.38

500

99.81

99.85

99.84

1,000

99.86

99.91

99.9

5,000

99.91

99.94

99.94

10,000

99.91

99.95

99.95

alpha

   

0.01

99.94

99.99

99.99

0.03

99.93

99.97

99.95

0.07

99.93

99.97

99.92

0.1

99.93

99.97

99.91

0.3

99.91

99.96

99.95

0.5

99.91

99.95

99.95

  1. The admixed population was generated from the African (YRI) and European (CEU) population from HapMap III. A sample from an admixed population is known to consist of a mosaic of chromosomal regions or blocks from the ancestral population. With increasing time since the admixture event, these regions are becoming broken up into smaller pieces through recombination that is denoted by the number of break points (Nbreaks). Individuals from the synthetically admixed population were sampled randomly from blocks from source populations, respectively (the defined admixture proportions, alpha). Finally, a subsample (nsnps) of uniformly distributed sites was chosen. The distance of the sites has been chosen to be greater than 1 Mb to assure linkage equilibrium.