Skip to main content

View full-text article in PMC

. 2019 Jan 22;10:379. doi: 10.1038/s41467-018-08222-6

Table 3.

Force field parameters of ML models optimized using our developed workflow

ML-BOP
m ^a	Gamma ^a	λ₃ (Å⁻¹) ^a
1.0	1.0	0.0
C	d	cosθ₀	n	β
77638.534354	16.148387	−0.471029	0.770018	1e-06
λ₂ (Å⁻¹)	B (eV)	R (Å)	D (Å)	λ₁ (Å⁻¹)	A (eV)
2.199640	473.621419	3.282761	0.270511	2.750522	1684.301476
ML-BOP _dih

m ^a	Gamma ^a	λ₃ (Å⁻¹) ^a
1.0	1.0	0.0
C	d	cosθ₀	n	β
77638.534354	16.148387	−0.471029	0.770018	1e-06
λ₂ (Å⁻¹)	B (eV)	R (Å)	D (Å)	λ₁ (Å⁻¹)	A (eV)
2.199640	473.621419	3.282761	0.270511	2.750522	1684.301476
cosθ_1a	cosθ_1b	cosθ_2a	cosθ_2b	k_dih (eV)	p
0.156434	0.017452	−0.390731	−0.5	0.2e-3	8
ML-mW ^b

ϵ (eV)	σ (Å)	a	λ	γ	cosθ₀
0.297284	1.884015	2.124872	24.673877	1.207943	−0.279667
A	B	p	q	tol ^a
7.111598	1.991526	4.011214	0.0	0.0

^aParameters that are not optimized in our ML workflow

^bSee Supplementary Equation 1–3 for the functional form (Stillinger-Weber, same as mW¹⁰)