Prepare MD trajectories for Sassena
-----------------------------------

Sassena requires a pdb file and a dcd file that define the trajectory.
The pdb file has to contain one single time frame of the system and is only used to
assign an element to each atom. The dcd file then contains the whole trajectory.

One has to convert the trajectories to pdb/dcd, for example using the program VMD.

There are usually two options to store trajectories: "wrapped" and "unwrapped".

Wrapped means that the atoms' coordinates are always within the simulation box. In case of periodic boundary boxes, this means that the atoms will jump to the opposite end if they transverse the box side. These trajectories keep the correct density, but the dynamics of these jumping atoms is completely artificial. Since Sassena has no information about the periodicity of the system, it takes these jumps at face value and will output a wrong dynamics.

In unwrapped trajectories, the atoms are not placed back into the simulation box. If they perform long-range motions, they will move further and further away from their starting point. The density of the system will gradually decrease, but the simulation will not produce artefacts in the dynamics.

Computations related to the structure / pair correlations (diffraction patterns) should be run on wrapped coordinates.

Computations related to the dynamics / self correlations (incoherent intermediate scattering functions) should be run on unwrapped coordinates.

If you are interested in the coherent intermediate scattering function, you are in a bit of a pickle: Since you are interested in pair correlations, you want the correct density and therefore the wrapped trajectory. However, since you are interested in the dynamics, you want the correct dynamics and therefore the unwrapped trajectory.