Alexandria Chemistry Toolkit

Contents:

• 1. Preface
  • 1.1. Before you begin…
  • 1.2. About the Alexandria Chemistry Toolkit
  • 1.3. Motivation for the ACT
• 2. User Manual
  • 2.1. Installation of the ACT
    • 2.1.1. Prerequisites
    • 2.1.2. Conda Environment
    • 2.1.3. Running the Installation
    • 2.1.4. Troubleshooting Installation
    • 2.1.5. Testing the ACT
  • 2.2. Using the ACT
    • 2.2.1. ACT File formats
    • 2.2.2. Creating a new force field file from scratch
    • 2.2.3. Train your first force field
    • 2.2.4. Is my training hitting a wall?
    • 2.2.5. Running training using parallel processing
  • 2.3. MD Simulations with Alexandria Force Fields
    • 2.3.1. Using ACT simulate
    • 2.3.2. The ACT-OpenMM interface
      • 2.3.2.1. MD simulations using OpenMM
      • 2.3.2.2. Energy Minimization using OpenMM
  • 2.4. Predicting Molecular Properties
    • 2.4.1. Electrostatic Potential
    • 2.4.2. Electrostatic Moments
    • 2.4.3. Polarizability
    • 2.4.4. Normal Mode Analysis
      • 2.4.4.1. Infrared Spectra
      • 2.4.4.2. Thermochemistry
    • 2.4.5. Second Virial Coefficient
• 3. Reference Manual
  • 3.1. Energy Function
    • 3.1.1. Non-bonded interactions
      • 3.1.1.1. Coulomb Interaction using Point charges
      • 3.1.1.2. Coulomb Interaction using Gaussian charges
      • 3.1.1.3. Coulomb Interaction using Slater charges
      • 3.1.1.4. Long range Coulomb interactions
      • 3.1.1.5. Polarization
      • 3.1.1.6. Induction correction
      • 3.1.1.7. Pauli Repulsion and London Dispersion
      • 3.1.1.8. Treatment of long range dispersion
      • 3.1.1.9. Combination Rules for Van der Waals Potentials
    • 3.1.2. Bonded interactions
      • 3.1.2.1. Harmonic potential
      • 3.1.2.2. Morse potential
      • 3.1.2.3. Wei-Hua potential
      • 3.1.2.4. Angle potential
      • 3.1.2.5. Angle potential for linear compounds
      • 3.1.2.6. Out-of-plane vibrations
      • 3.1.2.7. Torsion potential
      • 3.1.2.8. Proper dihedral
    • 3.1.3. Special potentials
    • 3.1.4. Virtual Sites
    • 3.1.5. Total energy
  • 3.2. Force Field Training Algorithms
    • 3.2.1. MCMC
      • 3.2.1.1. Metropolis Criterion
      • 3.2.1.2. Multiple MCMC runs
    • 3.2.2. Genetic Algorithm
      • 3.2.2.1. Initialization
      • 3.2.2.2. Deviation from data
      • 3.2.2.3. Sorting
      • 3.2.2.4. Penalties
      • 3.2.2.5. Selection probabilities
      • 3.2.2.6. Rank
      • 3.2.2.7. Fitness
      • 3.2.2.8. Boltzmann
      • 3.2.2.9. Elitism
      • 3.2.2.10. Selection
      • 3.2.2.11. Crossover
      • 3.2.2.12. Mutation
      • 3.2.2.13. Termination
    • 3.2.3. HYBRID
  • 3.3. Force field design
    • 3.3.1. Physical Background
    • 3.3.2. Determining atom types
    • 3.3.3. Determining partial charges
    • 3.3.4. Charge Equilibration with Shells or Virtual Sites
  • 3.4. Force Field Training Targets
    • 3.4.1. Algorithm to compute energy components in ACT
    • 3.4.2. Monomer Energies and Forces
    • 3.4.3. Other Properties
  • 3.5. Parameters for Force Field Training
  • 3.6. Training Data
    • 3.6.1. Using existing data
    • 3.6.2. Alexandria Library
    • 3.6.3. Donchev data set
    • 3.6.4. Non-covalent interaction atlas
    • 3.6.5. ACT data
    • 3.6.6. Generating SAPT data
    • 3.6.7. Generating single molecule data
    • 3.6.8. Conversion to ACT molprop files
• 4. Interactive Notebooks
  • 4.1. Charge Methods
    • 4.1.1. Electronegativity equalization method
    • 4.1.2. Split charge equilibration
  • 4.2. Atom Typing using RDKit
• 5. Developer Manual
• 6. References