novemberff 0.6.1

Custom implementation of the AMBER forcefield, focused on simplicity and verbose outputs.

NOVEMBERFF

NovemberFF (implemeNtatiOn of VErbose aMBER ForceField) is a custom implementation of the AMBER forcefield, inspired by OpenMM's C++ source. NovemberFF's aim is not to be performant, but rather to be compact and easy to modify/customize. It also provides by default the option to obtain verbose outputs, i.e. the energies of every singular molecular interaction in arrays (instead of their sum).

• Current outputs supported:

  • Bonded energies
  • Angle energies
  • Dihedral (Torsion) energies
  • LennardJones energies
  • Coulomb energies

• Current forcefield formats supported:

  • XML

• Current forcefields packed along NovemberFF (at novemberff/_data):

  • RNA.OL3.xml for RNA
  • amber99sb.xml for protein
  • amber99sb-ucb.xml for protein

Requirements

• currently depends on MDAnalysis for parsing the PDB structures

Examples

• display energy sums: python3 main.py rna testdata/1ato.pdb python3 main.py prot testdata/prot.pdb

• save energy arrays: python3 main.py rna testdata/1ato.pdb testdata/output/1ato python3 main.py prot testdata/prot.pdb testdata/output/prot

TODO

• improve CLI.
• add documentation.
• add more forcefields.
• improve performance.
  • remove MDAnalysis dependency for PDB parsing.
  • improve bond_graphs's graph data structure.
• improve the residue/atom labels fixing procedures.
• finish refactoring of parser base class and forcefield.from_xml (i.e. should be Parser's responsability to yield the needed data, not Forcefield).
• allow to specify the forcefield to use in main.py.
• improve the option for saving the molecule's topology features (bonds, angles, etc), as well as their geometry values. Add it to the CLI. Add tests.
• allow the XML parser to deal with files were new lines are missing.
• improve Utils.solve_forcefield_path