Table 5 GDT loss predicted by SAAP-QA for the CASP12 blind test set of stage 2 models of 17 target full length strutures.

From: Role of solvent accessibility for aggregation-prone patches in protein folding

Stage 2 (150 Models) - Full Length Protein

Targets

Best GDT

Top10 GDT Loss

Top 5 GDT Loss

Top 1 GDT Loss

T0893

61.98

0.52

0.52

4.86

T0895

72.92

0.00

0.00

0.00

T0902

50.75

2.42

2.92

3.17

T0906

91.90

1.05

1.96

4.05

T0910

87.91

3.28

3.81

5.15

T0911

65.99

0.74

3.98

6.62

T0912

47.79

5.47

5.47

6.05

T0913

66.57

0.37

0.37

5.55

T0917

84.18

4.90

4.90

4.90

T0920

50.42

4.45

4.45

8.29

T0921

70.65

0.18

0.18

0.18

T0928

63.27

0.59

0.59

11.44

T0943

60.61

8.55

8.94

14.10

T0944

74.11

1.19

3.36

12.85

T0946

45.21

0.00

0.09

2.41

T0947

66.43

2.00

8.57

8.57

T0948

76.68

3.02

3.02

5.21

Average GDT Loss

2.28

3.12

6.08

  1. GDT loss corresponding to the difference between the best model available in the pool of decoys and the model captured by SAAP-QA, are listed for top 10, top 5 and top 1 category models. These full protein targets correspond to the 18 target domains listed in Table 3.
Back to article page