Table 3

Prediction Results for three test sets.

No. of genes for prediction

T-test (P-value)

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

Youden index


56 gene model


FH2

56

0.011

70.0

84.6

77.8

78.6

0.55

FH3

56

0.010

80.0

100

100

0.83

0.8

ATHRO1

56

0.019

71.4

71.4

71.4

71.4

0.43


COXEN model


48

0.0024

70.0

92.3

87.5

80.0

0.62

6

1.5e-10

100

100.0

100.0

100

1.00

25

0.010

78.6

66.7

68.8

76.9

0.45


Data sets FH2, FH3, and ATHERO1 were used to evaluate the classification performance of our atherosclerosis prediction model. We tested if the predicted risk of atherosclerosis could statistically discriminate familial hypercholesterolemia or general atherosclerosis patients with normal people in these test sets. The statistical significance of the difference in predicted scores (COXEN SCORE) was evaluated using the two-sample t-test between patients and normal groups. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were also calculated.

Cheng et al. BMC Medical Genomics 2012 5:2   doi:10.1186/1755-8794-5-2

Open Data