I'm interested in doing molecular docking. What are some free programs for it and what are their pros and cons?
Pros: Rigid and flexible receptors, highly customizable
Cons: Need to get free charmm, creating scripts can sometimes be laborious, no high performance GPU modules
Autodock Vina has flexible ligands by default, although selected (or all) dihedrals (a.k.a. torsions) can be made rigid by the user. The receptor is rigid by default, but flexible regions can be selected by the user.
PLANTS uses a class of stochastic optimization algorithms called ant colony optimization (ACO).
ACO-based search engine two scoring functions (PLANTS_CHEMPLP and PLANTS_PLP) flexible protein side-chains rigid-body docking of multiconformer libraries into rigid and flexible receptors constraint system docking with selected explicit, displaceable water molecules fully automatic ligand setup (rotatable bond identification, atom typing ...) virtual screening rescoring capability
Note: the docking software GOLD uses and recommend the PLANTS score function.
1. Ab-initio protein--protein docking:
GRAMM (=Global Range Molecular Modeling)
- Performs rigid-body docking based on the shape complementarity and Miyazawa-Jernigan contact potentials;
- Exhaustively samples the search space (i.e., identifies the Global Energy Minimum solution, in the given formulation);
- Very fast (employs Fast Fourier Transform) and is suitable for high-throughput docking;
- Allows clustering of docking poses as a postprocessing;
- currently no account for flexibility of molecules;
- most likely will fail in application to antibodies;
- is effectively a set of protocols/pipelines that cover a wide range of docking problems. There is no single universally applicable protocol to all cases.
- Efficient application requires analysis of the target and choosing the most suitable pipeline.
- An inherently flexible docking protocol that employs Normal Modes Analysis at both the global docking stage (ATTRACT) and the refinement stage (iATTRACT).
- quite good benchmarking results for protein--protein complexes (in terms of the CAPRI criteria)
- The package is written in a mixture of C, Python, bash and FORTRAN and is difficult to navigate/modify;
- virtually impossible to prepare input to the pipeline. For this purpose the developers created a web-server specifically for the preparation of input and configuration files that must be downloaded and run on a locally compiled code.
2. Template-based protein--protein docking:
There is a class of cases when one has an experimentally-solved structure of a protein--protein complex with monomers very similar in structure to those that must be docked: i.e.
A-B is an experimentally-solved structure of a dimer complex (monomers A and B)
one has monomers A' and B' (the target complex A'-B') such, that TM-score(A,A') > 0.5 and TM-score(B, B') > 0.5. Then, A' might be structurally aligned onto A and B' might be structurally aligned onto B. Thus, one gets the structure of complex A'-B'.
- robust and easy.
- templates (i.e., A-B) must be known.
Discussion of these issues and the latest benchmarking results can be found here:
Pros: Peptides are fully flexible. Makes use of Rosetta's excellent protein modeling facilities.
Cons: Requires a Rosetta installation (free). Doesn't yet support disulfide bridges or cyclization.
Patchdock is a Molecular Docking Algorithm Based on Shape Complementarity Principles.
- Web Based
- Over 100 possible docked structures with the first one being the best structure.
- Takes over 2.5 to 3 hours for the results to be sent to you via mail
Open Drug Discovery Toolkit
Open Drug Discovery Toolkit provides nice features of rescoring with different scoring function your docked poses. This is important due to the limitations of scoring functions.