Table 1. Interaction Energy Error Statistics of the 1UBQ Fragment Database.

Method	μAll	σ2 All	μVdW	σ2 VdW	μPolar	σ2 Polar	R-factora
GAFF	0.36	3.26	0.25	0.36	0.46	5.64	0.127
FF99SBb	0.73	4.04	0.12	1.27	1.22	5.83	0.170
FF03b	0.83	6.61	0.18	0.81	1.36	10.86	0.259
AM1	3.15	9.50	1.04	0.70	4.85	10.28	0.373
PM3	2.65	7.89	0.14	0.77	4.67	4.59	0.352
PM6	1.67	2.24	0.84	0.32	2.34	2.82	0.211
PM6-DH2	0.30	1.23	−0.09	0.10	0.62	1.95	0.071
PDDG	3.21	16.23	−0.62	0.90	6.30	7.48	0.484
HF/6-31G*	1.94	1.32	2.27	1.14	1.68	1.30	0.153
HF/aDZ	2.14	1.22	2.29	1.11	2.02	1.29	0.176
HF/aTZ	2.10	1.17	2.28	1.10	1.95	1.17	0.171
HF/aQZ	2.08	1.16	2.28	1.10	1.93	1.15	0.170
MP2/6-31G*	1.24	0.64	1.12	0.28	1.34	0.91	0.146
MP2/aDZ	0.48	0.16	0.21	0.01	0.69	0.19	0.061
MP2/aTZ	0.16	0.02	0.05	0.00	0.24	0.02	0.023
B97-D/TZVP	0.20	1.06	−0.29	0.02	0.60	1.58	0.087
M06/6-31G*	0.75	0.42	0.63	0.12	0.85	0.64	0.104
M06/aTZ	0.73	0.16	0.57	0.08	0.85	0.19	0.090
M06-L/6-31G*	0.71	0.43	0.40	0.10	0.96	0.57	0.103
M06-L/aTZ	0.75	0.14	0.55	0.07	0.91	0.14	0.096

Mean and variance of interaction energy deviations (kcal/mol) from reference energies (a mix of MP2/CBS and CCSD(T)/CBS) for the interacting fragment molecules present in ubiquitin. The set of fragments was divided into 42 van der Waals interactions and 50 polar interactions. The related plots are presented in Table S1. a) The calculated R-factor serves as an analogy to the residual minimized in crystallographic structure refinement. A desirable value of R-factor would be less than 0.1. b) The force field- based atomic charge parameters were scaled to yield correct net charge on each fragment system.