We studied the conformational dynamics of our protein model
by means of extended MD-simulations covering a total
of 232 nanoseconds.
Using the well-known Verlet-algorithm with an integration step size of
one femtosecond,
232 simulations, each of one nanosecond duration, were carried
out
.
All 232 simulations started with almost identical initial conditions,
derived from the structure resulting from the
folding procedure described in Section 
by
minute random modifications of atomic
positions (
).
Nevertheless, these 232 simulations are
essentially independent from each other, since
the chaoticity inherent in the dynamics guarantees
a rapid decorrelation of the initially similar trajectories within
a few picoseconds.
An integration time step size as short as one femtosecond may seem to contradict the purpose of our simplified model, i.e., the reduction of the computational effort that has to be spent to study conformational dynamics. Note, however, that here the main computational speed up is due to the decrease of the number of degrees of freedom as compared to more realistic protein models. Furthermore, as described in the preceding Chapter, conformational transitions in our model are expected to be accelerated, implying fewer integration steps per conformational transition. As we shall see below, these expectations will be confirmed, in that the simulations actually will exhibit a large number of conformational transitions within the time span of our simulations.