Table 1.

Performance of the final regressor on a non-redundant blind test set. The model was trained including both forward and hypothetical reverse mutations, and tested on a non-redundant blind test set. The performances of other methods are also shown for comparison purposes.

Name	Method	Pearson	Spearman	Kendall	RMSE (Kcal/mol)
SUSPECT-ABL	Machine Learning	0.74	0.57	0.42	0.40
mCSM-lig [49]	Machine Learning	0.43	0.33	0.24	0.75
Aldeghi et al.[11]	Machine Learning	0.06	0.21	0.15	0.64
Hauser et al.[9]	Molecular Dynamics	0.65	0.38	0.30	0.69
A99 [11]	Molecular Dynamics	0.40	0.26	0.17	0.68
A99L	Molecular Dynamics	0.60	0.37	0.26	0.55
A99DC	Molecular Dynamics	0.58	0.45	0.31	0.55
A14	Molecular Dynamics	0.37	0.37	0.26	0.70
R15[11]	Rosetta	0.72	0.50	0.38	0.49
R16	Rosetta	0.47	0.29	0.20	0.70