Overview of BAC usage models
Binding affinity calculator (BAC) is a tool facilitating the calculation of protein-ligand binding free energies using ensemble molecular dynamics simulations. It supports two methods of free energy calculation; enhanced sampling of molecular dynamics with the approximation of continuum solvent (ESMACS) and thermodynamic integration with enhanced sampling (TIES). In ESMACS continuum approximations and coarse grained representatiopns are used to calculate absolute binding affinities of a ligands to its binding partner. TIES is a more rigorous approach but is generally limited to computing relative binding free energies between two ligands.
Two different groups of users are supported by BAC. Expert modellers, whose knowledge allows them to build and assess starting structures and drug force field descriptions, and application focussed users, who execute studies on particular protein-ligand combinations.
The current model of BAC usage is based around an expert user who provides a starting model for each system to be studied. This starting model consists not only of a complete PDB of the complex, containing coordinates for all of the residues in the target protein and the ligand under investiagtion, but forcefield parameterizations for non-standard elements and information about residues protonation and bonding. Once this initial model is produced it can be used as the basis of a variety of studies conducted using the fully automated BAC workflow (see Figure X).
The parameterization used for the ligand and the protein model are retained and may form the basis of future studies (for instance with point mutations introduced to the protein structure). For each system of interest the expert user builds up a library of protein models and drug parameterizations using there knowledge of the system to provide chemical details not present in the source PDB.
Figure 1: BAC workflow.
Figure 1 shows the interaction between the inputs defined by the expert user (shown in light blue) and the workflow which has already been fully automated ( shown in red). In addition to the existing BAC builder, simulation and analysis workflow, we have developed a number of tools to support the creation of the initial model by the expert user (the inputs ringed in dark blue). These are focussed around drug parameterization (both of the individual drugs used in ESMACS and the hybrid topologies for TIES).
Expert users can control all aspects of the system to be studied. Decisions that may have to be made include whether to use the full protein or only a fragment of interest and the origin of the structure to be used. It is possible to initiate simulations from not only experimental structures (e.g. those derived from from x-ray crystallography or NMR) but also from homology models. Ligands may also have to be docked into structures. Furthermore, expert users may also wish to alter simulation parameters (such as the length of equilibration or the size and shape of water box employed) in order to find the best protocol for a given system.
Users without the necessary modelling skills to build their own systems work with pre-prepared structures and drug parameters produced by expert users. Their usage of BAC will be confined to the fully automated pipeline shown in the red box in Figure 1. The choices available to application users will be limited to the number of replicate simulations and whether point mutations are to be incorporated into the model.
- Drug parameter builder - Status: An Antechamber wrapper exists
- Hybrid topology builder - Status: Basic tool exists
- PDB scanner to evaluate preparedness for simulation - Status: A tool has been developed in CCP-SAS for CHARMM
- PDB fixing tools - Status: Need to evaluate existing tools and provide a wrapper
Other usage cases to be considered in future:
- Need more automated drug building
- Integrate with docking
- Run with all defaults