Table 2. Assessment of Surrogate Methods in the Low-Energy Regime ΔẼ ≤ 10 kcal/mola.
| cis |
trans |
|||||
|---|---|---|---|---|---|---|
| Surrogate | MAE
|
Av bb-RMSD | NLM | MAE
|
Av bb-RMSD | NLM |
| GAFF AM1-BCC | 7.1 ± 3.1 | 0.48 ± 0.25 | 10 | 5.9 ± 4.3 | 0.41 ± 0.12 | 15 |
| GAFF HF/6-31G(d) RESP | 6.1 ± 7.6 | 0.29 ± 0.25 | 10 | 6.1 ± 4.9 | 0.35 ± 0.17 | 29 |
| AMOEBApro13 | 3.6 ± 4.1 | 0.36 ± 0.11 | 23 | 4.8 ± 4.7 | 0.27 ± 0.16 | 49 |
| PM6 | 3.7 ± 4.6 | 0.37 ± 0.14 | 25 | 4.4 ± 3.4 | 0.23 ± 0.16 | 32 |
| PM7 | 5.5 ± 6.7 | 0.43 ± 0.18 | 31 | 2.6 ± 1.9 | 0.21 ± 0.06 | 13 |
| DFTB3-D3(BJ) | 6.9 ± 4.7 | 0.36 ± 0.21 | 34 | 6.0 ± 6.7 | 0.36 ± 0.22 | 45 |
| B3LYP/6-31G | 1.6 ± 2.8 | 0.16 ± 0.17 | 30 | 1.1 ± 0.9 | 0.13 ± 0.13 | 35 |
| B3LYP/6-31G(d,p), reference | 27 | 21 | ||||
NLM is the number of local minima within the
range.
in kcal/mol,
where
and
are
the respective cis or trans putative
GM found over the entire test set. The energies of the surrogate and
the reference are compared according to the smallest bb-RMSD match,
whose average value and standard deviation are reported in Å.

