Practical notes derived from the theorem of Molecular Dynamics

  • Force fields in YASARA
    • AMBER: 5 essential parts, good for explicit solvents no in vacuo
    • YAMBER-AMBER: AMBER adapted for yasara good for long protein simulations
    • YASARA/YASARA2: Best for homology modelling. uses dihedral potentials
    • YAMBER3:
      • What were using with a 7.86 anstrom cutoff for vdW forces only
      • long range electrostatic uses coulombs law via Wwalds algorithm
      • made for long simulations.
      • vdW interactions are a little stronger to eliminate holes in structure building
  • Prepare structure for simulation
    • Add hydrogens, solvate it, eliminate gps, eliminate steric clashes
    • Check amide sidechain orientation (Asn/Gln) because hydrogen is ambiguous
    • Check atom assignments of histidine assignments because Carbon vs Nitrogen is difficult
    • Final minimization
  • Dunk it in a box bound, usually water box
  • Some programs have to specify which cysteines are bridged
  • Initial state
    • position r is from experimental (crystallographers or QM)
    • velocities are random values in boltzman distribution modulated by temperature
  • Simulation
    • initial state r and v
    • sum of forces are vector F on each atom via force field
    • calculate accelerations from newtons 2nd law
    • use a small time interval dt
    • calculate dv = integral(a x dt)
    • new velocity vector v’ = v + dv
    • new positions r’ = r + v’ dt
    • take new r’ and v’ and return to step 2
  • Periodic boundary conditions
    • every time an atom exits from boundary conditions of the cell, an identical one of same type enters from opposite side with same velocity
  • Equilibration
    • Have to wait for the system to equilibrate at the given temperature because initial V are random
    • this means you can’t use the first part of the simulation
    • for protein you dump the first nanosecond
    • for peptide this doesn’t take as long, 100-300 picoseconds
    • you know when Heavy Atom root mean standard deviation (HA_rmsd) plotted over time should equilibrate
    • Energy should be pretty constant over time