explicit
atomistic detail (many degrees of freedom)
expensive (many molecules, solvent relaxation)
solute-solvent interactions
samples smaller portion of phase space
implicit
mean-field detail (PB equation, GB approach)
inexpensive (no solvent degrees of freedom)
misses some solute-solvent features (e.g. hydrogen bonds)
samples bigger portion of phase space
imagine two states, X and Y
similar to chemical kinetics
detailed balance
transition probability
equal to opposite move
Boltzmann distribution
metropolis choice (satisfies condition)
equivalent to
Henderson Hasselbach
MD people use protonation fraction
interactions cause deviations
protonated fraction in terms of free energy
by comparing with
we get
md samples configurations under constant protonation states
we want to sample configurations under constant pH
also we want titration curve to calculate pKa and Hill coefficient
hybrid method, leverages monte carlo and classical molecular dynamics
The key quantity is the free energy of protonation
(Amber
manual
p. 563)
no knowledge of
need to use reference compound
thermodynamic cycle
(Vinny's dissertation)
AMBER approximation
within a few lines of algebra
)
MC move
MD/MC will sample entire protonation space, ergodicity
get
multiple