Table 1. Prediction performance for WTCCC data on test data.
Disease/method | T1D | T2D | BD | CD | CAD | RA | HT |
BootRank + Majority | 0.90 | 0.82 | 0.83 | 0.70 | 0.72 | 0.74 | 0.68 |
GWASRank + Majority | 0.88 | 0.69 | 0.68 | 0.67 | 0.72 | 0.75 | 0.65 |
LO, AC [15] | 0.75 | 0.6 | 0.67 | 0.63 | 0.6 | 0.67 | 0.61 |
SVM [67] | 0.82 | 0.71 | |||||
GWASelect [24] | 0.79 | ||||||
SVM, LR [63] | 0.89 | ||||||
Forward ROC [64] | 0.71 | ||||||
LR, SVM, RF, BN [65] | 0.56 | ||||||
Elastic-net [16] | 0.64 | ||||||
LR, AC, SVM [66] | 0.6 |
Shown are the AUC values obtained by different studies across the seven diseases in the WTCCC dataset. The reported AUCs were calculated only for test individuals. For each study, we took the best AUC reported for each disease, and missing diseases were left blank.
Diseases: T1D, Type 1 diabetes; T2D, Type 2 diabetes; CD, Crohn's disease; CAD, coronary artery disease; BD, bipolar disorder; RA, rheumatoid arthritis; HT, hypertension. Algorithms: SVM, support vector machine; LR, logistic regression; AC, allele count; RF, random forest; LO, log odds; BN, Bayesian networks.