Topology generation modules

    Table of contents
    No headers

    Modules for generating topologies. All the modules should be callable from an integrated application, in such a way that one can optionally perform a calculation (energy minimization, MD etc.), starting from a pdb file without running any intermediate programs.


    We also need to remove all knowledge about specific molecules, atoms, particles, or interactions from Gromacs and have them in data files instead. To the extent possible, modules should this be generic, not protein-specific!


    • Name: Structure cleaning/building.
      Take a "dirty" structure and clean it up, rename atoms, generate missing atoms etc.
      EL110409: This will probably be a sequence of modules, one for each class of molecules.
      Modules for biomolecules:
      • Alternative Protonation    
        From a clean structure  compute the hydrogen bonding network (including solvent if present), and detect molecules that might have alternative protonations for one or more sites. Protonation could be set either by pH, P-B calculations, or from hydrogen bond networks in the initial conformation (in particular for pairwise alternative sites common e.g. in biomolecules). The code should not assume proteins, though.
      • Sidechain Building from rotamer libraries
        Not sure how generic we can make this one...
    • Name: Topology generation.
      Assuming a perfect input file (with exactly the hydrogens we want in the simulation), this module will apply knowledge from force field data files stored on disk to generate a topology object. No atoms should be added or removed to keep it simple. Virtual sites for hydrogens should likely not go in here, but be moved so we can use it as an setting (vsites=on/off) later.
      This will likely be a very large module that needs to be split further.
    • Name Force field handling.
      Read and write force field files.
    Page last modified 14:38, 9 Apr 2011 by lindahl