Biomolecules in the energy landscape framework

This page is intended to collect information how the potential energy landscape framework can be used for biomolecules. Some information is provided directly on this page, and longer tutorials are link from here.

Set up

The setup for biomolecular simulations depends on the force field used for the simulations. The three main options in the code at the moment are AMBER, CHARMM, and OPEP. The OPEP interface includes not only the OPEP interface but also the HiRE-RNA force field. As AMBER have developed a quite good interface for working with pdb files the LEAP suit (either through tleap or xleap) are good choices. Another good option is the PyMOL software - it is very useful for structural representations in publication quality, checking correct that pdb files are set up correctly, as well as docking biomolecules together.

AMBER

Within AMBER there are two options to choose from: AMBER9 and AMBER12. As the old force fields are still available through AMBER12 and a number of newer projects are exclusively written for AMBER12 it is probably worth to stick with this newer option. The setup for AMBER9 and AMBER12 is essentially identical, and good tutorial for AMBER9 can be found here. For AMBER12 the key differences are a larger choice of force fields and solvent models as well as the support for GPUs. For proteins you should use the ff14SB or the ff99SB force fields. If using ff14SB, there are two solvent models you can choose from (both supported using CUDA) - igb2 and igb8. The former uses set default PBradii mbondi2, while igb8 needs set default PBradii mbondi3 in Leap. igb8 is meant to be somewhat more accurate, but if used with RNA it can lead to Abnormal termination messages. If this happens, just try switching to igb2.

However, there are a few things that are important to be noticed. First of all, you need to change the libraries in your AMBER installation with the one provided in the SVN. The changes in the provided libraries are symmetrised charges for NPHE and NLEU. The new libraries are found in the AMBERTOOLS/dat/leap/lib directory, you will need to copy them to your AMBER version (AMBER12 or newer!). When creating the perm.allow file with capping groups errors can occur for NME and NMET for non-canonical pdb formats. It is best to check for such problems before running calculations.

OPEP

OPEP allows to use coarse graining for proteins as well as for RNA and DNA. It will be faster than AMBER but due to the coarse graining some features will not be represented well in the potential and the landscape exhibits very large flat regions, which make the complicate the location of transition states. A full guide to the OPEP interface and the necessary input files can be found here.

GMIN

To run GMIN you need an executable compiled with OPEP, CHARMM, AMBER9 or AMBER12 and potentially CUDA. Examples for using OPEP, CHARMM, and AMBER9 guide through the setup for these interfaces. The AMBER12 interface and the AMBER9 interface are again essentially similar. However, while for AMBER9 MD moves as well as group rotation moves exist, for AMBER12 only the latter can be used. For alternative schemes there are mutational steps, basin-sampling, and free energy basin-hopping

Group rotation moves

An additional input file is needed for using grouprotation moves. Several scripts exist to set up this framework:

• a fortran script at ~/svn/SCRIPTS/AMBER/rigidbody/groupRigidBodyINC.f90
• a python script allowing very individual setups at ~/svn/SCRIPTS/AMBER/rigidbody/genrigid_input.py
• a python script without flexibility (used for mutational steps) at ~/svn/SCRIPTS/AMBER/BHmutation_steps/grouprotations.py
• a python script for group rotation moves to be paired with rotamer moves at ~/svn/SCRIPTS/AMBER/rotamer_moves/atomgroup_gen.py

Rigid bodies

The use of rigid bodies can help to reduce the computational time significantly. An example of how to use FSA can be found here. A more recent script to set up the input for rigid bodies, which accepts proteins, RNA, DNA and mixed systems is described here.

Another way of constraining motion is given through the use of constraints within AMBER.

OPTIM

The use of OPTIM with AMBER9 is described here. The setup for AMBER12 is identical apart from the small changes required when the system is setup. Recently, QCI has been improved to allowed the location of complete initial paths for more complex motions.

PATHSAMPLE

Various ressources describe how PATHSAMPLE may be used for biomolecules. Good introductions are given here. A more specific introduction to AMBER12 and some problems encountred is here.

Scripts and analysis

There is a large number of methods and scripts that may help the analysis of your results. A few of the more important ones are:

• How to decode features of Cv curves
• Analyse the distribution of minima and transition state energies
• The largest connected component
• How to construct free energy disconnectivity graphs