Table 3.

Receiver Operating Characteristics (ROCs) Area for Seven Predictors Using the Varibench–dbSNP [Sherry et al., 2001; Sasidharan Nair and Vihinen, 2013] Dataset (n = 17,329)

	Confidence	ROC area	Standard error	Min ROC area	ax ROC area
UMD‐Predictor	0.95	0.828	0.004	0.821	0.836
SIFT	0.95	0.784	0.005	0.774	0.794
PPH2_VAR	0.95	0.826	0.004	0.819	0.834
PROVEAN	0.95	0.789	0.004	0.781	0.797
CONDEL	0.95	0.778	0.004	0.771	0.786
MUT‐ASS	0.95	0.813	0.004	0.806	0.820
CADD	0.95	0.834	0.004	0.827	0.841

Min ROC area, lower bound for the confidence interval of a vector of length two; Max ROC area, upper bound for the confidence interval of a vector of length two. All data were generated using the “ci.cvAUC” function of the “cvAUC” package (https://github.com/ledell/cvAUC) for the ROCR R‐package [Sing et al., 2005]. UMD‐Predictor values were obtained without the mutations’ frequency information.