Trajectory reconstruction - martaenciso/BIT_FED GitHub Wiki

Swarm trajectories run for very short periods of time and new ones are relaunched, hoping to explore the whole conformational space in a fast and thorough way. Keeping track of how a certain configuration has evolved along the different epochs is important in order to analyse the complete final results.

This task, called trajectory reconstruction in BIT_FED, is carried out inside the script BIT_FED_analysis.sh and is composed by the following steps:

1. Identification of parent snapshots: the script calls the FORTRAN program trajectory_reconstruction.f90, which reads the chosen_from_epochN.out files generated along the simulation run and identifies which snapshots from a previous epoch were used as parents for the subsequent epoch. Keep in mind that several copies of each parent are started in the following epoch (check the Simulation flow page in this wiki for further details). The snapshot sequence of all possible trajectories longer than a certain value (chosen by the user) is then stored in files called trajectory_t_c.txt, where t is the time length of the reconstructed trajectory (in picoseconds) and c is an identification number of that trajectory.

2. Creation of reconstructed files: once the sequences of events have been identified, the script cuts and pastes the appropriate information into reconstructed dcd files, that store the tridimensional trajectory files, and colvars.traj files, which hold the collective variables for each snapshot.

The reconstructed trajectories show how an initial configuration has evolved and explored different conformations along the BIT_FED run. Besides, they are used in subsequent analysis such as free energy calculations (insert link).