Table 1.

Comparison of the predictive performance of functionality prediction tools on pathogenic and pharmacogenetic data sets

Algorithm	Category	Performance on disease-associated data set (AUCROC)	Performance on pharmacogenetic data set (AUCROC)
SIFT	Functionality prediction algorithms	0.76–0.88	0.74
PolyPhen-2		0.79–0.88	0.77
LRT		0.67–0.72	0.75
MutationAssessor		0.8–0.83	0.78
FATHMM		0.87–0.91	0.51
FATHMM-MKL		0.91	0.73
PROVEAN		0.85	0.76
VEST3		0.91	0.8
GERP++	Evolutionary conservation scores	0.67–0.78	0.67
SiPhy		0.69–0.81	0.63
PhyloP (vertebrate)		0.67–0.83	0.64
PhyloP (mammalian)			0.64
PhastCons (vertebrate)		0.67–0.83	0.58
PhastCons (mammalian)			0.61
CADD	Ensemble scores	0.93	0.81
DANN		0.95	0.75
MetaSVM		0.88–0.89	0.68
MetaLR		0.92–0.94	0.68

Performance measures on disease-associated data sets were obtained from refs. [26–31]