Computational Estimation of Microsecond to Second Atomistic Folding Times

Upendra Adhikari, Barmak Mostofian, Jeremy Copperman, Sundar Raman Subramanian, Andrew A. Petersen, Daniel Zuckerman

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Despite the development of massively parallel computing hardware including inexpensive graphics processing units (GPUs), it has remained infeasible to simulate the folding of atomistic proteins at room temperature using conventional molecular dynamics (MD) beyond the microsecond scale. Here, we report the folding of atomistic, implicitly solvated protein systems with folding times τ ranging from ?10 μs to ?100 ms using the weighted ensemble (WE) strategy in combination with GPU computing. Starting from an initial structure or set of structures, WE organizes an ensemble of GPU-accelerated MD trajectory segments via intermittent pruning and replication events to generate statistically unbiased estimates of rate constants for rare events such as folding; no biasing forces are used. Although the variance among atomistic WE folding runs is significant, multiple independent runs are used to reduce and quantify statistical uncertainty. Folding times are estimated directly from WE probability flux and from history-augmented Markov analysis of the WE data. Three systems were examined: NTL9 at low solvent viscosity (yielding τ f = 0.8-9 μs), NTL9 at water-like viscosity (τ f = 0.2-2 ms), and Protein G at low viscosity (τ f = 3-200 ms). In all cases, the folding time, uncertainty, and ensemble properties could be estimated from WE simulation; for Protein G, this characterization required significantly less overall computing than would be required to observe a single folding event with conventional MD simulations. Our results suggest that the use and calibration of force fields and solvent models for precise estimation of kinetic quantities is becoming feasible.

Original languageEnglish (US)
Pages (from-to)6519-6526
Number of pages8
JournalJournal of the American Chemical Society
Volume141
Issue number16
DOIs
StatePublished - Apr 24 2019

Fingerprint

Molecular Dynamics Simulation
Viscosity
Proteins
Molecular dynamics
Uncertainty
Protein Folding
Calibration
Parallel processing systems
History
Rate constants
Temperature
Water
Trajectories
Fluxes
Hardware
Kinetics
Computer simulation
Graphics processing unit

ASJC Scopus subject areas

  • Catalysis
  • Chemistry(all)
  • Biochemistry
  • Colloid and Surface Chemistry

Cite this

Adhikari, U., Mostofian, B., Copperman, J., Subramanian, S. R., Petersen, A. A., & Zuckerman, D. (2019). Computational Estimation of Microsecond to Second Atomistic Folding Times. Journal of the American Chemical Society, 141(16), 6519-6526. https://doi.org/10.1021/jacs.8b10735

Computational Estimation of Microsecond to Second Atomistic Folding Times. / Adhikari, Upendra; Mostofian, Barmak; Copperman, Jeremy; Subramanian, Sundar Raman; Petersen, Andrew A.; Zuckerman, Daniel.

In: Journal of the American Chemical Society, Vol. 141, No. 16, 24.04.2019, p. 6519-6526.

Research output: Contribution to journalArticle

Adhikari, U, Mostofian, B, Copperman, J, Subramanian, SR, Petersen, AA & Zuckerman, D 2019, 'Computational Estimation of Microsecond to Second Atomistic Folding Times', Journal of the American Chemical Society, vol. 141, no. 16, pp. 6519-6526. https://doi.org/10.1021/jacs.8b10735
Adhikari, Upendra ; Mostofian, Barmak ; Copperman, Jeremy ; Subramanian, Sundar Raman ; Petersen, Andrew A. ; Zuckerman, Daniel. / Computational Estimation of Microsecond to Second Atomistic Folding Times. In: Journal of the American Chemical Society. 2019 ; Vol. 141, No. 16. pp. 6519-6526.
@article{9e2b2d82403d44159b6e58b92d31316d,
title = "Computational Estimation of Microsecond to Second Atomistic Folding Times",
abstract = "Despite the development of massively parallel computing hardware including inexpensive graphics processing units (GPUs), it has remained infeasible to simulate the folding of atomistic proteins at room temperature using conventional molecular dynamics (MD) beyond the microsecond scale. Here, we report the folding of atomistic, implicitly solvated protein systems with folding times τ ranging from ?10 μs to ?100 ms using the weighted ensemble (WE) strategy in combination with GPU computing. Starting from an initial structure or set of structures, WE organizes an ensemble of GPU-accelerated MD trajectory segments via intermittent pruning and replication events to generate statistically unbiased estimates of rate constants for rare events such as folding; no biasing forces are used. Although the variance among atomistic WE folding runs is significant, multiple independent runs are used to reduce and quantify statistical uncertainty. Folding times are estimated directly from WE probability flux and from history-augmented Markov analysis of the WE data. Three systems were examined: NTL9 at low solvent viscosity (yielding τ f = 0.8-9 μs), NTL9 at water-like viscosity (τ f = 0.2-2 ms), and Protein G at low viscosity (τ f = 3-200 ms). In all cases, the folding time, uncertainty, and ensemble properties could be estimated from WE simulation; for Protein G, this characterization required significantly less overall computing than would be required to observe a single folding event with conventional MD simulations. Our results suggest that the use and calibration of force fields and solvent models for precise estimation of kinetic quantities is becoming feasible.",
author = "Upendra Adhikari and Barmak Mostofian and Jeremy Copperman and Subramanian, {Sundar Raman} and Petersen, {Andrew A.} and Daniel Zuckerman",
year = "2019",
month = "4",
day = "24",
doi = "10.1021/jacs.8b10735",
language = "English (US)",
volume = "141",
pages = "6519--6526",
journal = "Journal of the American Chemical Society",
issn = "0002-7863",
publisher = "American Chemical Society",
number = "16",

}

TY - JOUR

T1 - Computational Estimation of Microsecond to Second Atomistic Folding Times

AU - Adhikari, Upendra

AU - Mostofian, Barmak

AU - Copperman, Jeremy

AU - Subramanian, Sundar Raman

AU - Petersen, Andrew A.

AU - Zuckerman, Daniel

PY - 2019/4/24

Y1 - 2019/4/24

N2 - Despite the development of massively parallel computing hardware including inexpensive graphics processing units (GPUs), it has remained infeasible to simulate the folding of atomistic proteins at room temperature using conventional molecular dynamics (MD) beyond the microsecond scale. Here, we report the folding of atomistic, implicitly solvated protein systems with folding times τ ranging from ?10 μs to ?100 ms using the weighted ensemble (WE) strategy in combination with GPU computing. Starting from an initial structure or set of structures, WE organizes an ensemble of GPU-accelerated MD trajectory segments via intermittent pruning and replication events to generate statistically unbiased estimates of rate constants for rare events such as folding; no biasing forces are used. Although the variance among atomistic WE folding runs is significant, multiple independent runs are used to reduce and quantify statistical uncertainty. Folding times are estimated directly from WE probability flux and from history-augmented Markov analysis of the WE data. Three systems were examined: NTL9 at low solvent viscosity (yielding τ f = 0.8-9 μs), NTL9 at water-like viscosity (τ f = 0.2-2 ms), and Protein G at low viscosity (τ f = 3-200 ms). In all cases, the folding time, uncertainty, and ensemble properties could be estimated from WE simulation; for Protein G, this characterization required significantly less overall computing than would be required to observe a single folding event with conventional MD simulations. Our results suggest that the use and calibration of force fields and solvent models for precise estimation of kinetic quantities is becoming feasible.

AB - Despite the development of massively parallel computing hardware including inexpensive graphics processing units (GPUs), it has remained infeasible to simulate the folding of atomistic proteins at room temperature using conventional molecular dynamics (MD) beyond the microsecond scale. Here, we report the folding of atomistic, implicitly solvated protein systems with folding times τ ranging from ?10 μs to ?100 ms using the weighted ensemble (WE) strategy in combination with GPU computing. Starting from an initial structure or set of structures, WE organizes an ensemble of GPU-accelerated MD trajectory segments via intermittent pruning and replication events to generate statistically unbiased estimates of rate constants for rare events such as folding; no biasing forces are used. Although the variance among atomistic WE folding runs is significant, multiple independent runs are used to reduce and quantify statistical uncertainty. Folding times are estimated directly from WE probability flux and from history-augmented Markov analysis of the WE data. Three systems were examined: NTL9 at low solvent viscosity (yielding τ f = 0.8-9 μs), NTL9 at water-like viscosity (τ f = 0.2-2 ms), and Protein G at low viscosity (τ f = 3-200 ms). In all cases, the folding time, uncertainty, and ensemble properties could be estimated from WE simulation; for Protein G, this characterization required significantly less overall computing than would be required to observe a single folding event with conventional MD simulations. Our results suggest that the use and calibration of force fields and solvent models for precise estimation of kinetic quantities is becoming feasible.

UR - http://www.scopus.com/inward/record.url?scp=85064987021&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064987021&partnerID=8YFLogxK

U2 - 10.1021/jacs.8b10735

DO - 10.1021/jacs.8b10735

M3 - Article

VL - 141

SP - 6519

EP - 6526

JO - Journal of the American Chemical Society

JF - Journal of the American Chemical Society

SN - 0002-7863

IS - 16

ER -