Combined QM/MM Modeling Methods

Interactions in the QM/MM coupling

$$E_{total} = E_{QM} + E_{MM} + E_{QM/MM}$$ $$ E_{QM/MM} = E_{ES}(QM/MM) + E_{vdW}(QM/MM) + E_{bonded}(QM/MM) $$

The effective Hamiltonian

$$\hat{\mathcal{H}}_{eff} = \hat{\mathcal{H}}_{QM} + \hat{\mathcal{H}}_{ES}(QM/MM)$$

 

 




Average Solvent Potential in Mean-Field Theory

The mutual solute–solvent polarization

The ASEP/MD Hamiltonian $$ \hat{\mathcal{H}} = \hat{\mathcal{H}}_{QM} + \hat{\mathcal{H}}_{class} + \hat{\mathcal{H}}_{int} $$ The effective Schrodinger equation $$ (\hat{\mathcal{H}}_{QM}+\hat{\mathcal{H}}_{int}) \left | \Psi \right \rangle = E \left | \Psi \right \rangle $$ The interaction Hamiltonian associated to the electrostatic and van der Waals contributions $$ \hat{\mathcal{H}}_{int} = \hat{\mathcal{H}}^{elect}_{int} + \hat{\mathcal{H}}^{vdw}_{int} $$ The MFA electrostatic interaction(a statistical average over configurations) $$ \left \langle \hat{\mathcal{H}}^{elect}_{int} \right \rangle = \int {\hat{\rho} \cdot \left \langle V_{S}(r) \right \rangle } \, dr $$ - $\hat{\rho}$ : the solute charge density operator
- $\left \langle V_{S}(r) \right \rangle$ : the average electrostatic potential by the solvent

The effective Schrodinger equation considering the MFA energy $$ (\hat{\mathcal{H}}_{QM} + \left \langle \hat{\mathcal{H}}_{int} \right \rangle) \left | \Psi \right \rangle = E \left | \Psi \right \rangle $$

Location of critical points on free energy surfaces

The force on the Helmholtz free energy surface $$ G = -k_{B}T\ln{Z_{NVT}} $$ $$ \left \langle F(R) \right \rangle = - \frac{\partial G(R)}{\partial R} = - \left \langle \frac{\partial E}{\partial R} \right \rangle = - \left \langle \frac{\partial E_{QM}}{\partial R} \right \rangle - \left \langle \frac{\partial E_{int}}{\partial R} \right \rangle $$

Calculation of free energy differences

Within the ASEP/MD methodology, the free energy difference in solution between two given states $$ \Delta G_{s} = \Delta E_{solute} + \Delta G_{int} + \Delta ZPE_{solute} $$ $$ \Delta E_{solute} = E_{B} - E_{A} = \left \langle \Psi_{B} | \hat{\mathcal{H}}^{0}_{B} | \Psi_{B} \right \rangle - \left \langle \Psi_{A} | \hat{\mathcal{H}}^{0}_{A} | \Psi_{A} \right \rangle $$




The Electronic Structure

Statistical mechanics sampling for many-body interacting systems in condensed phases

The Hamiltonian of the system $$\hat{\mathcal{H}} = -\frac{1}{2}\sum_{A=1}^{M}{\frac{1}{M_{A}}\nabla^{2}_{M}} -\frac{1}{2}\sum_{i=1}^{N}{\nabla^{2}_{N}} + V(\vec{r_{i}};\vec{R_{A}},\vec{X})$$

Electronic spectra in QM/MM

A QM solute(M) and MM solvent molecules(I) system

The QM/MM coupling schemes $$E_{QM/MM}[M] = E[M] = E_{QM}(M) + E_{MM}(\{ I \}) + E_{int}$$ $$E_{QM}(M) = \left \langle \Psi(\vec{r};\vec{R};\vec{X})| \hat{\mathcal{H}} | \Psi(\vec{r};\vec{R};\vec{X}) \right \rangle$$ Mechanical coupling $$E_{int}(el) = \sum_{A\alpha}\frac{q_{A}q_{\alpha}}{r_{A\alpha}} $$ Electrostatic coupling : Mechanical coupling with QM polarization considering potential for the solvent $$E_{int}(el) = \left \langle \Psi(r_{i};R_{A};q_{\alpha}) | \sum_{\alpha}\frac{q_{\alpha}}{r_{i\alpha}} | \Psi(r_{i};R_{A};q_{\alpha}) \right \rangle + \sum_{A\alpha}\frac{q_{A}q_{\alpha}}{r_{A\alpha}} $$ The wavefunctions for both ground and excited states $$\begin{align*} \Psi^{0} &= \Phi^{0}_{(0)} + \Phi^{0}_{(1)} + \Phi^{0}_{(2)} + \cdots \\ \Psi^{*} &= \Phi^{*}_{(0)} + \Phi^{*}_{(1)} + \Phi^{*}_{(2)} + \cdots \\ \end{align*}$$ QM Expectation for the excitation energy $$\begin{align*} \omega_{(0)} &= \left \langle \Phi^{*}_{(0)} | \hat{\mathcal{H}} | \Phi^{*}_{(0)} \right \rangle - \left \langle \Phi^{0}_{(0)} | \hat{\mathcal{H}} | \Phi^{0}_{(0)} \right \rangle \\ \omega_{(1)} &= \left \langle \Phi^{*}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{*}_{(0)} \right \rangle - \left \langle \Phi^{0}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{0}_{(0)} \right \rangle \\ \omega_{(2)} &= \left \langle \Phi^{*}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{*}_{(1)} \right \rangle - \left \langle \Phi^{0}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{0}_{(1)} \right \rangle \\ &\vdots \\ \omega_{(n)} &= \left \langle \Phi^{*}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{*}_{(n-1)} \right \rangle - \left \langle \Phi^{0}_{(0)} | \frac{q_{\alpha}}{r_{i\alpha}} | \Phi^{0}_{(n-1)} \right \rangle \\ \end{align*}$$

Many body system expansion

The QM/MM coupling schemes $$E_{QM/MM} = E[M] = \left \langle \Psi | \hat{\mathcal{H}} +\sum_{\alpha}\frac{q_{\alpha}}{r_{i}r_{\alpha}} | \Psi \right \rangle + E_{MM}(solvent) + E_{int}(vdW) $$ The total system energy expression making a distinction between the solute M and the solvent molecules {I} $$E = E[M] + \sum_{I \neq M}{E[I]} + \sum_{I \neq M}{(E[MI]-E[M]-E[I])} + \sum_{I>J\ \& \ i,j \neq M}{(E[IJ]-E[I]-E[J])}$$ The energy difference between the ground($M^{0}$) and a specific excited($M^{*}$) state $$\begin{align*} \Delta E = &E[M^{*}] - E[M^{0}] \\ &+ \sum_{I \neq M}{\Delta E[I]} \\ &+ \sum_{I \neq M}{(E[M^{*}I]-E[M^{0}I] -E[M^{*}]+E[M^{0}] -\Delta E[I])} \\ &+ \sum_{I>J\ \& \ i,j \neq M}{(\Delta E[IJ] -\Delta E[I] -\Delta E[J])} \\ \end{align*}$$ $$\text{where } \Delta E[I] = E[I]_{M^{*}} - E[I]_{M^{0}}$$

QM methods for the calculation of electronic spectra



SCC-DFTB : The self-consistent charge-density functional tight-binding method

Density functional tight-binding method



Self-consistent charge–density functional tight-binding



A posteriori treatment for London dispersion in SCC–DFTB






The Effect of the Protein Environment

Born–Oppenheimer approximation



Conical intersections



Excited state molecular dynamics



Diabatic surface hopping



Mixed quantum classical molecular dynamics

The potential energy of the system by MM force-field $$ V_{MM} = \sum_{i}^{N_{bonds}}V^{bond}_{i} + \sum_{j}^{N_{angles}} V^{angle}_{j} + \sum_{l}^{N_{torsions}} V^{torsion}_{l} + \sum_{i}^{N_{MM}}\sum_{j>i}^{N_{MM}} V^{Coulomb}_{ij} + \sum_{i}^{N_{MM}}\sum_{j>i}^{N_{MM}} V^{LJ}_{ij} $$




QM/MM Methodology and Applications

Energy Expression Schemes

  • Subtractive scheme
  • Additive scheme

Free Energy Perturbation(FEP)

$$\Delta F(A\rightarrow B) = -k_{B}T\ln{\left < e^{-\frac{E_{B}-E_{A}}{k_{B}T}} \right >_{A}}$$

Thermodynamic Integration(TI)

Potential energy function $$U(\lambda) = \lambda U_{A} + (1-\lambda)U_{B}$$ The partition function of the system in the canonical ensemble $$Q(N,V,T,\lambda) = \sum_{s} e^{-\frac{-U_{s}(\lambda)}{k_{B}T}}$$ The free energy of this system $$F(N,V,T,\lambda) = -k_{B}T\ln{Q(N,V,T,\lambda)}$$ The ensemble average of the derivative of potential energy with respect to $\lambda$ $$\Delta F(A\rightarrow B) = \int_{0}^{1} {\frac{\partial F(\lambda)}{\partial \lambda}}\, d\lambda = \int_{0}^{1} {\frac{\partial U(\lambda)}{\partial \lambda}}\, d\lambda$$

Enhanced sampling

  • Multiple time-step sampling
  • Umbrella sampling
  • Replica Exchange
  • Reaction coordinate-driven methods
  • Transition path sampling
 

 

 


Reference
  1. Theoretical Concepts for Chemical Energy Conversion
  2. QM/MM Methods for Biomolecular Systems
Related Articles
  1. [Chemistry] Ab initio methods
  2. [Chemistry] Density Functional Theory
  3. [Chemistry] Molecular Mechanics and Dynamics
  4. [Chemistry] Quantum Chemistry Methods