Molecular Dynamics

Molecular dynamics (MD) is a simulation method that computes the trajectories of atoms and molecules by numerically solving Newton's equations of motion for a system of interacting particles. Unlike machine learning predictors that infer structures from statistical patterns, MD attempts to reproduce the actual physical process by which biomolecules move, fold, and interact — treating the molecule as a dynamical system rather than a static object.

The method is computationally expensive: a single microsecond of protein motion may require days of supercomputer time, creating a scale gap between what can be simulated and what can be experimentally observed. This has driven interest in coarse-grained models that sacrifice atomic detail for temporal reach, and in machine-learned force fields that accelerate simulation while retaining physical fidelity. Whether either approach preserves the causal structure of the phenomena they model is an open question in computational biophysics.