Skip to main content

Table 2 Performance comparison of Bustard, Ibis and AYB on several sets of reads of varying read length and chemistry versions.

From: All Your Base: a fast and accurate probabilistic approach to base calling

  

Reads mapped, %

  

Reads perfect, %

 
 

Bustard

Ibis

Δ%

AYB

Δ%

Bustard

Ibis

Δ%

AYB

Δ%

Ï•X174 L2

76.62

78.33

+2.23

78.25

+2.13

55.88

58.94

+5.48

62.29

+11.48

Ï•X174 L4

63.02

66.11

+4.90

65.09

+3.29

40.09

43.08

+7.46

44.74

+11.60

Ï•X174 L6

72.09

74.07

+2.75

74.08

+2.77

51.19

53.34

+4.20

56.00

+9.40

Ibis Test

84.77

88.45

+4.34

88.19

+4.03

44.34

66.14

+49.16

69.32

+56.34

B. pert./1

28.76

39.16

+35.94

45.80

+58.98

2.53

3.14

+23.70

4.13

+62.86

... trimmed

77.35

81.06

+4.80

81.14

+4.90

39.52

47.64

+20.55

55.24

+39.79

B. pert./2

34.33

47.41

+38.75

53.50

+55.57

6.22

17.69

+183.98

26.67

+327.97

... trimmed

66.54

70.22

+5.53

72.07

+8.31

30.13

40.72

+35.18

48.25

+60.15

BGI/1

87.41

89.01

+1.82

88.85

+1.64

59.62

68.39

+14.70

69.29

+16.22

BGI/2

84.58

86.29

+2.03

86.52

+2.29

55.95

61.90

+10.64

63.30

+13.14

Illumina/1

97.58

97.80

+0.22

97.85

+0.28

72.55

75.80

+4.49

76.70

+5.73

Illumina/2

96.29

96.73

+0.46

96.82

+0.55

70.61

73.88

+4.63

74.66

+5.74

HiSeq/1

84.97

85.24

+0.32

85.97

+1.18

60.29

62.55

+3.75

64.50

+6.98

HiSeq/2

79.78

†

 

81.34

+1.76

49.79

†

 

55.58

+11.63

  1. Performance is compared in terms of the percentage of reads mapped back to the reference with five edits or fewer, and the percentage of reads which perfectly match the reference; the 'Δ%' figures for Ibis and AYB, where given, are the percentage improvements over Bustard. See text for further details. †Ibis failed to process the second end of the HiSeq data.