I arrived at New Orleans on Saturday and had the pleasure to watch very good presentations at the Permeation and Transport Subgroup session.
Molecular dynamics (MD) simulations of any system involve some approximations, like the use of classical mechanics instead of quantum mechanics to describe the time evolution of the system, or the representation of atoms with fixed point charges. The lectures of prof. Jana Shen and prof. Sergei Noskov (link) at the Permeation and Transport Subgroup session on Saturday employed methods to overcome some of these approximations.
In conventional MD simulations, the protonation states of titratable groups are attributed based on pKa values obtained from experimental conditions that usually do not resemble the conditions of these groups in the protein structure. Moreover, once assigned the protonation state is fixed for the whole simulation. Prof. Jana Shen highlighted that such approximations can lead to incorrect protonation state assignments, neglect of coexistence of the deprotonated and protonated states for a given pH and fail to offer mechanistic insights where changes in the protonation state matter. To avoid these approximations, prof. Jana Shen used in her work continuous constant pH MD (CpHMD) simulations, a type of simulation where the protonation state of a titratable group changes according to the environment. Using this method, she was able to reveal molecular details of the mechanism of Na+ transport in the NhaA antiporter.
Prof. Sergei Noskov showed in the beginning of his presentation that the presence of a divalent ion inside the ion channel pore can modify the electronic structure up to the second solvation shell. This result was obtained by quantum mechanical calculations. In conventional MD simulations, on the other hand, atom charges are fixed, and polarizability effects as those caused by the divalent ion would not be accounted for. To include polarizability in MD simulations, prof. Sergei Noskov employed a drude model, parametrized by him and co-workers, to describe his system. In the drude model, each atom is represented by two charged particles attached, what allows some change in the charge distribution of a single atom. He showed some pathways for ion permeation in NavAb and CavAb channels , and that some stable configurations identified by MD simulations with the drude model were missed by classical MD simulations.
MD simulations have many approximations, and whether we need to avoid them or not depend on the question we want to answer. The two lecturers above wanted to answer questions that required improvement of some of the current approximations, and they were able to employ methods for higher quality modeling to answer their questions successfully.