Question 1

How is adequate sampling demonstrated?

Accepted Answer

Length alone proves nothing, since a trajectory can remain in one basin indefinitely. Several independent runs from different starting velocities, agreement between them on the measured property, block averaging with error estimates, and analysis showing the relevant coordinate has been crossed repeatedly together make the case. Reporting per-replicate values rather than a pooled average is expected.

Question 2

How much does force field choice affect conclusions?

Accepted Answer

Considerably, and unevenly. Folded globular regions are reasonably robust across modern parameter sets, while disordered segments, ion binding, and lipid or nucleic acid environments differ markedly. Where a result depends on one of these, repeating with a second force field and reporting both outcomes is a far stronger position than a single choice defended in passing.

Question 3

Which setup details must be documented?

Accepted Answer

Protonation states assigned at the working pH, especially for histidine and titratable residues near the site of interest; water model; ion type and concentration; box dimensions and periodic treatment; cutoff scheme and long-range electrostatics method; thermostat and barostat; and the equilibration sequence. Omitting any of these makes the work impossible to reproduce.

Question 4

When are enhanced sampling methods appropriate?

Accepted Answer

When the transition of interest is too slow for unbiased simulation, which covers most conformational changes worth studying. Choosing collective variables that actually capture the transition is the difficult part, and a poor choice yields confident but wrong free energy surfaces. Convergence should be assessed by reweighting and by independent runs with different starting points.

Question 5

How should binding free energy results be reported?

Accepted Answer

With uncertainty derived from independent repeats rather than from a single run, along with convergence diagnostics for each window or lambda point where relevant. Correlation with experimental affinities across a series matters more than agreement on one ligand, and the method, protocol and any restraints should be stated in full.

Question 6

What experimental comparisons anchor a simulation?

Accepted Answer

Order parameters from relaxation measurements, chemical shifts, distance distributions from spectroscopic labelling, hydrogen exchange rates, and scattering profiles can all be computed from trajectories and compared directly. Even partial agreement on one observable gives a simulation standing that internal consistency alone cannot provide.

Biotechnology

Biotechnology

Research Frontiers in Molecular Dynamics Simulation Refinement