1. Working and up to date install of python, ProDy, and networkx
   • pip install prody
   • pip install networkx
2. For do_minimization.py, a working installation of gromacs 4.5.

All trajectories should be in multi-frame PDB format.

cartesian_average.py trajectory.pdb average.pdb

Iteratively superimpose all frames of the trajectory onto the average and then output the cartesian average of all heavy atoms.

find_centroid.py trajectory.pdb centroid.pdb

Compute the average as for cartesian_average.py and then output the frame from the trajectory that is closest to the average.

network_average.py --cutoff 1.0 trajectory.pdb average.pdb

Compute the NetAvg for trajectory. A cutoff of 1.0 works best on our benchmark set.

do_minimization.py start.pdb target.pdb final.pdb

Apply position restraints to move start closer to target, while maintaining a physical structure. A working version of gromacs must be in your path for this step. Uses implicit solvent and amber99sb-ildn forcefield, but the forcefield shouldn't matter much as the atom positions are largely dicated by the average.

Other than issues with installtion, the two most likely causes of probelms is atom naming. These tools work with my benchmark tests, but I suspect there will be problems when using other data sets due to differences in heavy atom naming conventions.