PD is a fully functional molecular mechanics engine written by myself and Jon Rea as part of our graduate studies. The main site for this website is here.
Below is the main overview description for this project.

PD – A Molecular Mechanics Engine for Python

PD is a highly modular, cross-platform molecular mechanics suite for Python. The library design is focused on providing unprecedented flexibility, modularity and adaptability without compromising speed of execution. The core algorithms are written in C++ for maximum efficiency and make extensive use of object-oriented design patterns. All features are automatically wrapped into Python using SWIG, a widely used wrapper generator, circumventing the need for the developer to know anything about the exporting process while still allowing python to be used as the input script language. This allows the user to set up simple simulations quickly and easily while providing powerful language features for the more advanced user. All abstract concepts in molecular simulations such as force fields and simulation methods are represented by classes which interact with one another in a highly modular fashion, allowing the user to create new algorithms without coding a single line of C++. For the developer the object oriented design of the core library allows easy reuse of code and rapid development and testing of new algorithms hence accelerating scientific progress. Furthermore, it is possible to write separate modules which interact with the core library but are separately compiled and maintained and enable scientific groups to provide new algorithms to the code base without the need for code patches to the core library. At this early stage, the library already provides a large number of widely used features: Various molecular dynamics protocols (e.g. Replica Exchange Simulation) and Monte Carlo algorithms as well as methods for free-energy calculation, quasi-harmonic and normal mode analysis, have been implemented; Algorithms for homology modeling and docking are under development. Commonly used force fields AMBER, CHARMM and OPLS are currently supported, including periodic boundary simulations. Implicit solvation, such as surface area based models and Generalized-Born solvent models are also implemented.

Downloads

Presentation from the CCPB meeting 2007       Download

Features

General

• Written in C++, wrapped into Python
• Exposure into Python allows very simple as well as highly complex use of the features giving the user a maximum of flexibility when setting up simulations.
• The design allows classes to recombine into novel algorithms at runtime (i.e. from python).
• Interface generated automatically using SWIG: No need to write interface code after a new feature is added.
• Runtime linkage allows external modules. In this way Academic Groups and Commercial users can contribute functionality to the simulation community in an external way and under licensing separate from PD.

Forcefield functional form

• Bonded: Harmonic Bonds & Angles, Fourier & Harmonic Torsions & Impropers
• NonBonded: Lennard Jones Van der Waals, Simple Electrostatics. Energy and Force switching.
• Spaces: Periodic boundary conditions (Orthogonal)
• Solvation: Distance-dependent Dielectric Constant, Solvent Accessible Surface Area (SASA) based solvation, Generalised Born / Surface Area (GB/SA)
• Knowledge Based: Residue-Residue Interaction matrix based
• Restraints: Harmonic Cartesian, Distance & Torsion Restraints & various other custom forces.
• Other: Go forcefield, Softcore VdW Forcefield
• OpenMP allows parallelism of NonBonded calculations across the CPUs of Multicore/MultiCPU machines.

Forcefields

• AMBER: (ff94, ff96, ff99, ff03),
• CHARMM: (Charmm19, Charmm22),
• Other: OPLS/aa, Mizyawa & Jernigan, RAFT

Protocols

• Gradient based Minimisation (Steepest Descent, Conugate Gradients & Torsional)
• Monte Carlo, Monte Carlo with Minimiation including a large variety of Structure Perturbation (Move Types) including Rotamer Libraries & Fragment Insertions
• Conformational Space Annealing (CSA)
• Homology Modelling: Threading & Loop Builder
• Molecular Dynamics: Beeman , Verlet & Velocity Verlet Integrators, Andersen and Berendsen Thermostats
• Langevin Dynamics
• Replica Exchange Dynamics (REMD/REMC)
• Numerical Second Derivate Matrix (Hessian) Calculation and Normal Mode Analysis
• Quasi Harmonic Analysis
• Free Energy Calculations: Free Energy Pertubation , Thermodynamic Integration, Umbrella Sampling

Formats

• Input: PDB, TRA
• Output: PDB, DCD/PSF, TRA

Availablity

We are currently setting up a licence agreement under which PD will be made available in the near future. We plan that it will be made available to Academic Institutions for free or at a very low price.

If you are interested please send an email to mike.tyka@bris.ac.uk and we will notify you when the program will be available.

Mike Tyka’s Thesis Chapter 3: