TY  - JOUR
T1  - Reproducible Molecular-Dynamics Evidence Requires More Than Longer Trajectories, Larger Systems, and Visually Persuasive Conformational Movies That Obscure Uncertainty
A1  - Hans Müller
A1  - Anna Schmidt
A1  - Thomas Weber
JF  - International Journal of Pharmaceutical Research and Allied Sciences
JO  - Int J Pharm Res Allied Sci
SN  - 2277-3657
Y1  - 2025
VL  - 14
IS  - 1
DO  - 10.51847/zxBEdTo5M1
SP  - 15
EP  - 26
N2  - Molecular-dynamics simulation can reveal motions, interactions, and state transitions that are difficult to observe directly, but its evidential strength is frequently inferred from computational scale rather than demonstrated through reproducibility. Longer trajectories, larger systems, elaborate enhanced-sampling calculations, and visually persuasive conformational movies may increase descriptive richness without resolving uncertainty arising from system preparation, force-field selection, initial conditions, incomplete sampling, collective-variable construction, analytical flexibility, or disagreement with external evidence. This methodological framework article proposes the Molecular-Dynamics Evidential Reproducibility Framework, an original non-empirical synthesis that connects claim specification, system provenance, model adequacy, independent sampling, estimator-specific convergence, uncertainty decomposition, analytical robustness, negative controls, external validation, and bounded claim formation. The framework is organized as a gated reproducibility chain in which no single component can substitute for a missing evidential function elsewhere. A complementary Claim-Strength Matrix classifies simulation claims from visual illustration to externally validated mechanistic inference according to cumulative evidence from replication, convergence, robustness, and validation. Eight testable propositions are derived, including the non-substitution, estimator-first, replication, analysis-invariance, model-perturbation, visualization-boundary, reproducibility–validity distinction, and cumulative-claim propositions. Reporting requirements and governance recommendations are proposed for authors, reviewers, and journals. The framework does not prescribe universal trajectory durations, system sizes, replica numbers, or numerical convergence thresholds. It is intended to make uncertainty and decision boundaries visible, not to certify a simulation as correct. Prospective evaluation is required before the framework can be treated as a validated reporting or editorial standard.
UR  - https://ijpras.com/article/reproducible-molecular-dynamics-evidence-requires-more-than-longer-trajectories-larger-systems-and-nngtjnz8jddwfxq
ER  -