Performance of Web tools for predicting changes in protein stability caused by mutations

BMC Bioinformatics

Table 2 General results from the assessment on the full dataset of monomeric proteins

	PoPMuSiC	DynaMut	DUET	INPS-MD	MAESTROweb
Values calculated for mutations causing a ΔΔG > \|0.5\|kcal/mol
True negative	424	270	413	422	399
False positive	19	173	30	21	44
True positive	44	80	64	49	62
False negative	58	22	38	53	40
Accuracy	0.86	0.64	0.88	0.86	0.85
True negative rate (specificity)	0.96	0.61	0.93	0.95	0.90
True positive rate (sensitivity)	0.43	0.78	0.63	0.48	0.61
Positive predictive value (precision)	0.70	0.32	0.68	0.70	0.58
Negative predictive value	0.88	0.92	0.92	0.89	0.91
MCC	0.47	0.31	0.58	0.50	0.50
Values calculated for mutations causing a ΔΔG ≤\|0.5\|kcal/mol
True negative	101	58	87	93	75
False positive	17	60	31	25	43
True positive	27	50	33	25	43
False negative	52	29	46	54	36
Accuracy	0.65	0.55	0.61	0.60	0.60
True negative rate (specificity)	0.86	0.49	0.74	0.79	0.64
True positive rate (sensitivity)	0.34	0.63	0.42	0.32	0.54
Positive predictive value (precision)	0.61	0.45	0.52	0.50	0.50
Negative predictive value	0.66	0.67	0.65	0.63	0.68
MCC	0.23	0.12	0.16	0.12	0.18

True negative and true positive values have been considered as those predictions that correctly predicted a negative and a positive sign for destabilizing and stabilizing mutations, respectively

ISSN: 1471-2105