DDGun: an untrained method for the prediction of protein stability changes upon single and multiple point variations

Table 2 Performances of the sequence-based and structure-based baseline methods on single site variations data sets

Coefficients derived from	Method	VariBench 1564 variants	Broom et al. 605 variants	S2648 2648 variants	P53 42 variants	Myoglobin 134 variants
VariBench	DDGun	0.50, 1.71	0.52, 1.77	0.50, 1.40	0.70, 1.45	0.48, 1.20
VariBench	DDGun3D	0.54, 1.70	0.62, 1.68	0.57, 1.33	0.67, 1.54	0.57, 1.0
S2648	DDGun	0.50, 1.71	0.52 1.77	0.50 1.38	0.70 1.48	0.48 1.16
S2648	DDGun3D	0.54, 1.71	0.62 1.68	0.57 1.33	0.67 1.57	0.58 0.98
Broom et al.	DDGun	0.48, 1.73	0.52, 1.78	0.49, 1.42	0.71, 1.41	0.45, 1.29
Broom et al.	DDGun3D	0.54, 1.69	0.62, 1.66	0.57, 1.32	0.68, 1.51	0.56, 1.0
Average	DDGun	0.49, 1.72	0.52, 1.78	0.50, 1.4	0.7, 1.45	0.47, 1.21
Average	DDGun3D	0.54, 1.70	0.62, 1.67	0.57, 1.33	0.67, 1.55	0.57, 0.99

The Pearson correlation coefficient and the root mean squared error (RMSE) in kcal/mol are defined in section Methods

ISSN: 1471-2105