Uncertainty Quantification for Molecular Dynamics

Patrone, Paul N.; Dienstfrey, Andrew

doi:10.1002/9781119518068.ch3

Cited by 8 publications

(9 citation statements)

References 104 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Emulator based methods [14] and the test harness [110] are used to analyse sensitivity and uncertainty, to evaluate scientific software and to calibrate complex computer simulators. Practical recommendations have been made on validation and reproducibility in the field of scientific research in general [111] including molecular dynamics [5,39]. Uncorrelated data need to be collected in sufficient quantities; autocorrelation analyses can be used to better understand if a time series represents an equilibrated system [111], but caution must be exercised as longer autocorrelation times may be uncovered if other relevant free energy minima are revealed [45].…”

Section: Generating Actionable Predictionsmentioning

confidence: 99%

“…Computational predictions are increasingly being used to predict outcomes and provide recommendations in a variety of industrial and policy-making contexts. One of the central aims of UQ is to facilitate decision making [5]; this is enabled by providing probabilistic statements about quantities of interest, in a timely fashion, and helping decision makers to decide what actions to take that maximize the likelihood of the desired outcome. Ensemble methods for probabilistic weather prediction have become a routine part of the practice of weather forecasting [6].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Uncertainty quantification in classical molecular dynamics

Wan

Sinclair

Coveney

2021

Phil. Trans. R. Soc. A.

130

View full text Add to dashboard Cite

Molecular dynamics simulation is now a widespread approach for understanding complex systems on the atomistic scale. It finds applications from physics and chemistry to engineering, life and medical science. In the last decade, the approach has begun to advance from being a computer-based means of rationalizing experimental observations to producing apparently credible predictions for a number of real-world applications within industrial sectors such as advanced materials and drug discovery. However, key aspects concerning the reproducibility of the method have not kept pace with the speed of its uptake in the scientific community. Here, we present a discussion of uncertainty quantification for molecular dynamics simulation designed to endow the method with better error estimates that will enable it to be used to report actionable results. The approach adopted is a standard one in the field of uncertainty quantification, namely using ensemble methods, in which a sufficiently large number of replicas are run concurrently, from which reliable statistics can be extracted. Indeed, because molecular dynamics is intrinsically chaotic, the need to use ensemble methods is fundamental and holds regardless of the duration of the simulations performed. We discuss the approach and illustrate it in a range of applications from materials science to ligand–protein binding free energy estimation. This article is part of the theme issue ‘Reliability and reproducibility in computational science: implementing verification, validation and uncertainty quantification in silico ’.

show abstract

Section: Generating Actionable Predictionsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Uncertainty quantification in classical molecular dynamics

Wan

Sinclair

Coveney

2021

Phil. Trans. R. Soc. A.

130

View full text Add to dashboard Cite

show abstract

“…If you read this guide in its entirety before performing a simulation, you will have a much better sense of what constitutes (in our minds) a thoughtful simulation study. Thus, we strongly advise that readers review and understand the concepts presented here, as well as in related reviews [9, 12, 13]…”

Section: Pre-simulation “Sanity Checks” and Planning Tipsmentioning

confidence: 99%

Best Practices for Quantification of Uncertainty and Sampling Quality in Molecular Simulations [Article v1.0]

et al. 2019

Self Cite

View full text Add to dashboard Cite

The quantitative assessment of uncertainty and sampling quality is essential in molecular simulation. Many systems of interest are highly complex, often at the edge of current computational capabilities. Modelers must therefore analyze and communicate statistical uncertainties so that “consumers” of simulated data understand its significance and limitations. This article covers key analyses appropriate for trajectory data generated by conventional simulation methods such as molecular dynamics and (single Markov chain) Monte Carlo. It also provides guidance for analyzing some ‘enhanced’ sampling approaches. We do not discuss systematic errors arising, e.g., from inaccuracy in the chosen model or force field.

show abstract

“…characterized by the name of the model (1S, 2S, 2SF, 3S*, 3SF*) and the coarse-graining resolution (1,3,4,5,6), see [36] for a detailed description of the models. In Fig.…”

Section: Model Selection and Model Inadequacymentioning

confidence: 99%

“…In turn, uncertainty can be defined [4] as a "parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measurand ". Uncertainty quantification (UQ) is essential for building confidence in model predictions and helping model-based decisions [5]. Monitoring uncertainties in computational physics/chemistry has become a key issue [6], notably for multi-scale modeling [7].…”

Section: Introductionmentioning

confidence: 99%

Bayesian calibration of force fields for molecular simulations

Cailliez¹,

Pernot²,

Rizzi³

et al. 2020

Uncertainty Quantification in Multiscale Materials Modeling

View full text Add to dashboard Cite

Molecular simulations are one of the most prominent discovery tools in science and engineering, widely adopted in applications ranging from drug discovery to materials design. The fidelity of the predictions of molecular simulations hinges on the reliability and accuracy of the chosen force fields. Thus, the quantification of uncertainties associated with the form of force fields and their parameters is a fundamental part of molecular modeling. This aspect has been rather overlooked in the past but constitutes today an active field of research. This review focuses on the data driven, Bayesian calibration of force fields. The calibration of molecular force fields is a very challenging task in the presence of an ever increasing amount, type and scale of reference data. Bayesian data analysis provides an invaluable tool to address this challenge, notably in the presence of model inadequacy. It enables a rigorous definition of uncertainties for the force fields and allows for data driven, robust predictions through molecular simulations. This chapter reviews recent methodological and algorithmic developments in Bayesian calibration of force fields, and their application to atomic and molecular fluids. We discuss Bayesian approaches and their computational challenges followed by a description of target applications. We believe that the advances in Bayesian inference presented in this article would be valuable for a number of other fields in Science and Engineering, that are revolutionized by Data science.

show abstract

Uncertainty Quantification for Molecular Dynamics

Cited by 8 publications

References 104 publications

Uncertainty quantification in classical molecular dynamics

Uncertainty quantification in classical molecular dynamics

Best Practices for Quantification of Uncertainty and Sampling Quality in Molecular Simulations [Article v1.0]

Bayesian calibration of force fields for molecular simulations

Contact Info

Product

Resources

About