It’s my view that we must become statistical biophysicists. Why statistical? Because microscopic behaviors must be repeated zillions of times to create macroscopic effects. Can you help me shift the thinking in our community? See below for a collaboration opportunity.
Running a single MD simulation of any system really should never be an option. That’s the realm of storytelling – not science. (An exception to the no-single-runs rule is if the simulation is obviously many times longer than key timescales of interest, which is so rare, we can just ignore the possibility.) The dangers of single simulations are simply too great. Multiple runs provide you with that most beautiful thing: a measure of variance.
I am interested in collaborating on a Lessons Learned article on the limitations of molecular dynamics. We have a nice data set of multiple runs on an interesting system showing that 500 ns per run isn’t nearly enough sampling to distinguish the behavior of wild-type from a couple of mutants as we’d hoped to do. Do you have a similar ‘not worth publishing’ dataset? Help fight publication bias, and join me in this effort (email@example.com).
While we’re worrying about MD, I want to get you worrying that there may be even more things to worry about we typically realize. We all know about force fields, sampling, molecular crowding (potentially in phase-separated regions), and the need for physiological membranes and ions. So what about ATP? I’m not talking about ATP’s role in regulating intra-cellular phase separation, but rather ATP simply as a chemical modulator of the environment, like ions.
Why worry about ATP? For reference, a molecule present at 1M concentration is effectively on a grid with ~12 Å spacing. So 1mM corresponds to ~120 Å spacing. ATP’s cellular concentration is ~10 mM … and how big is your protein? Does your protein have polar or charged groups on its surface?
And what other molecules are present in the cell above mM concentration? I don’t know. But surely there are many instances, especially in local regions given the now accepted view of the cell as compartmentalized even outside typical organelles.