Could anyone guide me to analyze protein-protein interaction energy in a MD trajectory?
I'm simulating a dimer and want to see the interaction energy between the two monomers. Is it the same way as specifying the two chains (say residues 1-100 and in the other monomer residues 101-200) and then extract all the energy terms, add them to get the total interaction energy as a function of time?
If you used PME, there is no way to decompose the mesh term in a pairwise manner. You can decompose the short-range nonbonded energies, but not the long-range terms. Considering only the short-range energies is also not necessarily all that meaningful, in addition to the considerations Gabriel mentioned.
Whenever you have two molecules interacting you need to solve a thermodynamic cycle to accurately calculate the free energy of interaction between the two of them. This is because you need to take a lot of transformations into account. For example, the energy of interaction between your two proteins depend on how much each one needs to deform themselves, when compared to free dimers in solution, in order to interact with one another, or how many solvent molecules needs to be extruded out / have their degrees of freedom restrained in the interaction sites. So the problem really isn't so trivial as you migth at first think.
Fortunately a lot of smart people have given this problem a lot of thought and so there are a lot of ways to solve this problem for the free energy of interactions for a wide range of situations. The mos accurate methods I know (Thermodynamic Integration and Alchemic Perturbation) to solve that are much too computationally intensive to use on such a big system as a dimer of proteins subunits but there are simpler ways such as MM-GB/PBSA, which is widely used when estimating the interaction free energy between a protein and a small ligand. Since the theory is pretty much the same you could start by looking into this MM-GB/PBSA thing. Maybe you can use the MM-GB/PBSA right out of the box, but keep in mind that approximate methods usually work better when you deal with the system for which they were first idealized to solve. Maybe there is such a method specifically for protein-protein interaction.
If you used PME, there is no way to decompose the mesh term in a pairwise manner. You can decompose the short-range nonbonded energies, but not the long-range terms. Considering only the short-range energies is also not necessarily all that meaningful, in addition to the considerations Gabriel mentioned.
You can use www.chemosophia.com for calculation of Gibbs free energy, enthalpy, entropy at each point of the traectory of the 'protein-protein' interactions (find "Physicochemical properties" service after registration at the web site). Besides, you can estimate atom-atomic electron overlaps using "Fast integration of atomic basins(neglect of high order overlaps" software in "Large organic and elementorganic molecular systems" service. All these computations are available now on-line for free.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics