Skip to main content

Table 4 Comparison of the covarion model and MBL models with different number of components for three real datasets

From: Evaluation of the models handling heterotachy in phylogenetic inference

 

-LnL

AIC

BIC

CV

 

Animal dataset (5,000 sites and 20 species)

one-component

86468.5

86506.5

86630.3

82.1 ± 7.9

two-component

86302.7

86378.7

86626.4

37.8 ± 13.5

three-component

86222.7

86336.7

86708.2

47.9 ± 10.7

four-component

86167.6

86319.6

86814.9

69.0 ± 17.2

five-component

86126.8

86316.8

86936.0

82.2 ± 21.2

Six-component

86087.1

86315.1

87058.1

NC

covarion

86300.7

86340.7

86471.0

0

 

plastid dataset (3,754 sites and 22 species)

one-component

78225.2

78267.2

78398.0

75.3 ± 8.8

two-component

78056.4

78140.4

78402.1

34.2 ± 24.5

three-component

77996.7

78122.7

78515.2

49.8 ± 15.6

four-component

77925.8

78093.8

78617.2

60.3 ± 21.0

five-component

77926.2

78136.2

78790.4

72.4 ± 22.0

six-component

77900.4

78152.4

78937.5

NC

covarion

78070.9

78114.9

78252.0

0

 

mitochondrial mammal dataset (3,591 sites and 17 species)

one-component

44285.9

44317.9

44416.9

45.9 ± 3.7

two-component

44154.8

44218.8

44416.8

16.6 ± 7.5

three-component

44127.6

44223.6

44520.5

34.2 ± 12.3

four-component

44081.2

44209.2

44605.1

38.2 ± 15.4

five-component

44071.9

44231.9

44726.8

NC

six-component

44072.3

44264.3

44858.2

NC

covarion

44187.1

44222.1

44330.4

0

  1. For CV, standard deviation can be easily computed and is thus indicated.