I have performed structure prediction for disordered region (75 aa) of a protein and performed molecular dynamics for 10 nano seconds using relevant force field. How to obtain the most stable conformation or state from the performed MD results and a small epitope region of 10-mer is also available as crystal structure for the comparisons.
The most stable state corresponds to the minimum of free energy (either Helmholtz free energy in NVT or Gibbs free energy in NPT MD simulations.) To determine the free energy landscape, however, is a difficult problem, because if you see it as a function of all degrees of freedom of the system, then it becomes very difficult to calculate it and to visualize. Therefore, often the problem reduces to determining the so-called reaction coordinates (or collective coordinates), which will define a free energy landscape, where you can perform a minimum free energy search for obtaining the most stable state. Determining the collectives coordinates will depend on the system and problem, however some choices can be suggested, for example, backbone dihedral angles (Phi,Psi), root mean square deviation and radius of gyration, first two or three principal component analysis (PCA). Then, the free energy can be constructed as a function one, two or three of these reaction coordinates (not very often more than two) and perform a minimum search for determining the most stable state.
Intrinsically disordered proteins do not have prominent structures. I am actually surprised that you say there is a crystal structure for a part of that protein, unless that region is not part of the disordered region of the protein. Also, 10 ns is not long enough to capture the different conformations of a protein.
For a more general case, you can use clustering of the different conformations in the trajectory based on their structure and use the representative structure of the different clusters to present the prominent conformations of the proteins. Again, you need a longer simulation to use this protocol.
being $ gmx mdrun -s input.tpr -rerun configuration.pdb
The issue is that while I able to use a special forcefield, create a topology file, and a .gro file with $gmx pdb2gmx ... -water none", I am not able to create the .tpr file.
I had high hopes that by using grompp with this .mdp file, I could create the .tpr file, and then do an mdrun. But I'm getting a lot of errors which I don't understand with respect to Verlet Lists and incompatibility with the latest version of Gromacs 5.1.3.
It would be easy if I was doing a solvation. But I need to do this vacuum, and I'm surprised why that's so difficult. Is it possible to create a .tpr file without an .mdp file? My superior would like me to figure out a sequence of commands that works and then we can loop through several files, and do some energy statistics.
An empirical potential based quasicontinuum theory (EQT) is proposed to predict the structure, concentration, and various potential profiles of water in confined environments. EQT seamlessly unifies the continuum theory given by the Nernst-Planck equation and the atomistic theory governed by interatomic potentials. In particular, the interatomic po...