Non-bonded cutoffs

There are no fixed rules that determine how much larger the neighbour list cutoff should be than the non-bonded cutoff. Clearly there will be a trade-off between the size of the cutoff and the frequency at which the neighbour hst must be updated the larger the difference, the less frequently will the neighbour list have to be updated. There may also be storage implications if the list is too large. 

OPLS force field Monte Carlo simulation NB = Non-bonded cutoff distance in A... 

Leusen et al. studied the crystal packing of ephedrine with several phosphoric acid stereoisomers to verify whether a relationship between enthalpy of formation and separation of enantiomers via diastereomeric salt formation exists.They considered four different force fields and determined that the CHARMM program, as implemented in QUANTA, was the best for their application. Their decision was based mainly on the knowledge that non-bonded cutoff distances are larger in CHARMM. A larger cutoff distance is necessary to account for the interaction between different hydrophobic layers in the crystal (interlayer distances of 14—16 A). These authors were unable to quantitatively determine the validity of their hypothesis, although qualitative explanations were derived for their observations. 

A the non-bonding cutoff was 4.5 A. 20 ps of molecular dynamics was performed at 1000 K and the system was cooled down slowly, followed by final energy minimization. The final sequential simulated annealing simulation was implemented using a total of 3598 NOE constraints. 

---