A library for generating smooth paths from ab-inito MD trajectories. These paths might serve as input for an NEB calculation. The key component of this package is the ability to translate from cartesian coordinates to redundant internal coordinates (bond lengths, angles, dihedrals) and to go the other way, from redundant internal coordinates to an "optimal" set of cartesian coordinates.
- Select a set of redundant internal coordinates
- currently, we're using ALL pairwise distances, all of the angles between sets of three atoms, a-b-c, that actually get "bonded" during the trajectory, and all of the dihedral angles between sets of 4 atoms, a-b-c-d, that actually get "bonded" during the trajectory. I observed empirically that using ALL pairwise distances seems to work better than only using pairs corresponding to actually bonded atoms.
- Transform each frame in a trajectory into the internal coordinate space
- Apply smoothing filter to the time-series from each internal coordinate
degree of freedom independently. Currently, we're using a zero-delay
low pass Buttersworth filter. A digital convolution filter and a polynomial
fit smoother are also available in the package though.
- The cutoff frequency is the key variable for the algorithm.
- For each frame, find a set of xyz coordinates that are optimally consistent
with the smoothed internal coordinates.
- This is a computationally intensive optimization problem, even with analytic derivatives and Levenberg-Marquardt. Fortunately it's embarrassingly parallel over the frames in the trajectory, as each frame is optimized separately. The code is currently running this step across multiple nodes with MPI.
- Apply a very slight smoothing in the cartesian coordinate space, to correct for jitters in the trajectory caused by imperfections in the internal -> xyz optimization. This obviously is done after RMSD alignment.
There's some code in this library that might be of interest for other tasks, besides trajectory smoothing.
- Calculation of bond lengths, bond angles and dihedral angles from cartesian coordinates.
- Analytic derivatives of bond lengths, bond angles, and dihedral angles with respect to cartesian coordinates.
- Find a set of cartesian coordinates given redundant internal coordinates as input, via a least-squares optimization (Levenberg-Marquardt).
- Context managers that make MPI code in python much more elegant.
- Kabsch RMSD in numpy.
- python 2.6+ (tested with 2.7.3)
- numpy, 1.6+ (tested with 1.6.2)
- scipy, 0.10+ (tested with 0.10.1)
- mpi4py 1.3+ (tested with 1.3)
Open up the file client.py and take a look. Try running it, either serially
or with mpi
$ python client.py
$ mpirun -np 2 client.py
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.