Table 5.

Comparison with other methods.

Dataset	Q2	P[D]	S[D]	P[N]	S[N]	C	AUC	PM
SIFT	0.61	0.62	0.66	0.60	0.56	0.22	0.64	95
CHASM	0.80	0.85	0.73	0.76	0.87	0.60	0.88	100
SPF-All	0.88	0.88	0.87	0.87	0.88	0.75	0.94	100
SPF-Cancer	0.90	0.91	0.90	0.90	0.91	0.81	0.96	100

Overall accuracy (Q2), positive predictive value (P) Sensitivity, Correlation coefficient (C) and area under the ROC curve (AUC) are defined in Methods section. D (Disease) and N (Neutral) are respectively driver and passenger cancer variants. The latter have been generated by CHASM. PM is the percentage predicted variants for the Synthetic dataset. The confidence interval for Q2, C and AUC calculated on the two subsets is ≤0.01.