Skip to main content

Table 6 Simulation C, discrepancies between datasets and programs

From: Hybridization of mouse lemurs: different patterns under different ecological conditions

Combination

A

B

C

D

E

Program

NEW-

HYBRIDS

STRUCTURE

Both programs

Both programs

Both programs

Datasets

L+E

L+E

L

E

L+E

Efficiency Mg

1.000

1.000

1.000

1.000

1.000

Efficiency Mm

0.999

1.000

0.999

1.000

0.999

Efficiency F1

1.000

1.000

1.000

1.000

1.000

Efficiency F2

0.990

0.960

0.988

0.976

0.990

Efficiency Mg-Bx1

0.883

0.688

0.868

0.852

0.883

Efficiency Mm-Bx1

0.884

0.696

0.877

0.820

0.884

Efficiency Mg-Bx2

0.547

0.279

0.504

0.502

0.547

Efficiency Mm-Bx2

0.532

0.248

0.514

0.454

0.532

Accuracy Hyb

0.999

0.999

0.999

1.000

0.998

Accuracy Mg

0.946

0.905

0.941

0.939

0.946

Accuracy Mm

0.944

0.902

0.942

0.931

0.944

Discrepant between sets

325

426

  

715

Disc. true Hyb

97.8%

99.5%

  

98.7%

Discrepant between programs

  

1098

949

1363

Disc. true Hyb

  

99.4%

99.8%

99.3%

N Hyb obs. per set/pair sets

41-57

30-46

40-55

37-55

41-57

  1. We accepted all individuals as hybrids that were identified in at least one of two corresponding datasets (combinations A and B), by at least one program (combinations C and D) and by at least one program in at least one of two corresponding datasets (combination E). Datasets: L: local sets (n = 100), E: enlarged sets (n = 100), L+E: combined evidence from corresponding local and enlarged sets; Discrepant between sets: individuals identified as hybrids in only one of two corresponding datasets; Discrepant between programs: individuals identified as hybrids with only one of the two programs; Disc true Hyb: percentage of true hybrids among discrepant individuals in the line above; N Hyb obs. per set/pair sets: range of observed numbers of hybrids in 100 datasets or pairs of corresponding datasets. The true number of hybrids per set was 60. We disregarded the 20,000 additional purebred individuals comprised in the enlarged datasets exclusively for the calculation of efficiencies, accuracies and numbers of discrepant individuals. Only six of these individuals were misclassified as hybrids in combination C.