Reliability of Short Molecular Dynamics Runs as Screening Labels: Noise Ceilings, Time Dependence and Adaptive Replicate Allocation
5 affiliations
- 1Department of Applied Mathematics
- 2Escuela de Ingeniería Informática de Segovia, Universidad de Valladolid
- 3Department of Computer Science
- 4Universidad de Valladolid
- 5Independent Researcher
Abstract
High-throughput molecular dynamics (MD) produces the labels on which electrolyte screening and machine-learning models are built, but each label comes from one short, noisy run whose value still depends on the simulated time. We quantify their reliability, and how to spend a simulation budget given it, on HTP-MD, the largest public set of polymer electrolyte trajectories (6,137 runs of 5 ns and 133 of 50 ns). Lithium motion is diffusive in only 2 short and 5 long runs, the apparent diffusivity at a 5 ns lag is 1.95 times that at 25 ns, and the first 5 ns of a run are 1.63 times faster than its second half, biasing every short label upwards. From 94 systems simulated twice, a single 5 ns end-point label has reliability 0.47 (95% CI 0.31 to 0.63): no model can correlate with such labels much beyond about 0.69, a random forest on monomer fingerprints is indistinguishable from that ceiling (0.71), and roughly half of the run-to-run noise comes from the initial structure. A short run ranks widely different polymers well (Spearman 0.93) but not similar ones (0.68 within the screening set). In simulations with the measured noise, giving further runs only to candidates whose membership of the top set is uncertain matches four uniform replicates with 1.8 runs per candidate for the top 10% and 1.2 for the top 1%; a check on real runs is consistent but underpowered. A decision guide and all code are provided.
Related
- Accuracy, Reliability, and Efficiency of a MACE Machine-Learning Interatomic Potential for Aspirin: A Reproducible Benchmark for Molecular Simulation
- Accept or abstain? A pre-specified test of reliability gates calibrated on held-out chemistry for molecular property and reaction-yield models
- Determination of fundamental properties of nitrogen from first principles. III. Temperature and frequency dependence of the molecular polarizability and magnetic susceptibility
- Density dependence of retrieved self-broadened CO2 line intensities from requantized molecular dynamics simulations
- Runtime-steered ultrashort molecular dynamics enables million-pose protein–ligand screening
BibTeX
@misc{hernandogallego2026reliability,
title = {{Reliability of Short Molecular Dynamics Runs as Screening Labels: Noise Ceilings, Time Dependence and Adaptive Replicate Allocation}},
author = {Francisco Hernando-Gallego and Dimitar Vasilev-Vasilev and Diego García-Santos and Alfonso Campo-Camacho and Alberto Gómez-Gómez and Francisco Hernando Gallego},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15010895/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010895/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Reliability of Short Molecular Dynamics Runs as Screening Labels: Noise Ceilings, Time Dependence and Adaptive Replicate Allocation
AU - Francisco Hernando-Gallego
AU - Dimitar Vasilev-Vasilev
AU - Diego García-Santos
AU - Alfonso Campo-Camacho
AU - Alberto Gómez-Gómez
AU - Francisco Hernando Gallego
PY - 2026
DA - 2026/10/22
PB - ChemRxiv
DO - 10.26434/chemrxiv.15010895/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010895/v1
AB - High-throughput molecular dynamics (MD) produces the labels on which electrolyte screening and machine-learning models are built, but each label comes from one short, noisy run whose value still depends on the simulated time. We quantify their reliability, and how to spend a simulation budget given it, on HTP-MD, the largest public set of polymer electrolyte trajectories (6,137 runs of 5 ns and 133 of 50 ns). Lithium motion is diffusive in only 2 short and 5 long runs, the apparent diffusivity at a 5 ns lag is 1.95 times that at 25 ns, and the first 5 ns of a run are 1.63 times faster than its second half, biasing every short label upwards. From 94 systems simulated twice, a single 5 ns end-point label has reliability 0.47 (95% CI 0.31 to 0.63): no model can correlate with such labels much beyond about 0.69, a random forest on monomer fingerprints is indistinguishable from that ceiling (0.71), and roughly half of the run-to-run noise comes from the initial structure. A short run ranks widely different polymers well (Spearman 0.93) but not similar ones (0.68 within the screening set). In simulations with the measured noise, giving further runs only to candidates whose membership of the top set is uncertain matches four uniform replicates with 1.8 runs per candidate for the top 10% and 1.2 for the top 1%; a check on real runs is consistent but underpowered. A decision guide and all code are provided.
ER -