Construction of gene clusters resembling genetic causal mechanisms for common complex disease with an application to young-onset hypertension

Table 1 Classification accuracy after establishing a cut-off percentage of risky combinatory genotypes in each gene cluster

Datasets	Population	Classification accuracy S1(S2)	Sensitivity S1(S2)	Specificity S1(S2)	Subjects with risky genotypes
					HT	NC
FHS_Affy500k (training)	Caucasian	99.7%⁺(99.1%)	0.99(0.98)	1.0(0.996)	87	239
Taiwan_Affy500k (training)	Taiwanese	98.2%⁺(93.4%)	0.97(0.97)	1.0(0.89)	97	70
WTCCC_Affy500k (test)	Caucasian	61.6%(62.9%)	0.25(0.26)	0.85(0.87)	214	336
Taiwan_Affy100k & Taiwan_Illu550k (test)	Taiwanese	98.2%(98.2%)	0.97(0.94)	1.0(1.0)	138^*	89*
Overall	Both	82.8%(82.5%)	0.68(0.69)	0.93(0.93)	536	734
Overall except WTCCC_Affy500k	Both	98.9%(97.5%)	0.98(0.97)	1.0(0.98)	322	398

S1: threshold is computed for each dataset; S2: threshold is computed for each platform; ⁺Classification accuracy of the dataset evaluated using a five-fold validation procedure exhibit similar result and is presented in Additional file 1: Table S5; *see Additional file 1: Figure S1 for detailed calculations.

ISSN: 1471-2164