The algorithm for denoising point clouds of annular forgings based on Grassmann manifold and density clustering DOI
Yucun Zhang, An Wang,

Tao Kong

et al.

Measurement Science and Technology, Journal Year: 2024, Volume and Issue: 35(11), P. 115004 - 115004

Published: July 24, 2024

Abstract In the industrial sector, annular forgings serve as critical load-bearing components in mechanical equipment. During production process, precise measurement of dimensional parameters is paramount importance to ensure their quality and safety. However, owing influence environment, manufacturing process can introduce varying degrees noise, resulting inaccurate measurements. Therefore, researching methods for three-dimensional point cloud data eliminate noise forging clouds significant improving accuracy This paper presents a denoising approach based on Grassmann manifold density clustering (GDAD). First, within manifold, core points are determined using parameters. Second, performed with Cauchy distance replacing Euclidean reduce impact outliers analysis results. Finally, search tree model was constructed filter out incorrect clusters. The fusion results achieved data. Simulation experiments demonstrate that GDAD effectively eliminates edge performs well point-cloud models levels intensity.

Language: Английский

Exploring Kinase Asp-Phe-Gly (DFG) Loop Conformational Stability with AlphaFold2-RAVE DOI
Bodhi P. Vani, Akashnathan Aranganathan, Pratyush Tiwary

et al.

Journal of Chemical Information and Modeling, Journal Year: 2023, Volume and Issue: 64(7), P. 2789 - 2797

Published: Nov. 20, 2023

Kinases compose one of the largest fractions human proteome, and their misfunction is implicated in many diseases, particular, cancers. The ubiquitousness structural similarities kinases make specific effective drug design difficult. In conformational variability due to evolutionarily conserved Asp-Phe-Gly (DFG) motif adopting out conformations relative stabilities thereof are key structure-based for ATP competitive drugs. These extremely sensitive small changes sequence provide an important problem sampling method development. Since invention AlphaFold2, world has noticeably changed. spite it being limited crystal-like structure prediction, several methods have also leveraged its underlying architecture improve dynamics enhanced ensembles, including AlphaFold2-RAVE. Here, we extend AlphaFold2-RAVE apply a set kinases: wild type DDR1 three mutants with single point mutations that known behave drastically differently. We show able efficiently recover stability using transferable learned order parameters potentials, thereby supplementing AlphaFold2 as tool exploration Boltzmann-weighted protein (Meller, A.; Bhakat, S.; Solieva, Bowman, G. R. Accelerating Cryptic Pocket Discovery Using AlphaFold. J. Chem. Theory Comput. 2023, 19, 4355–4363).

Language: Английский

Citations

24
Acceleration of Molecular Simulations by Parametric Time-Lagged tSNE Metadynamics DOI Creative Commons
Helena Hradiská,

Martin Kurečka,

Jan Beránek

et al.

The Journal of Physical Chemistry B, Journal Year: 2024, Volume and Issue: 128(4), P. 903 - 913

Published: Jan. 18, 2024

The potential of molecular simulations is limited by their computational costs. There often a need to accelerate using some the enhanced sampling methods. Metadynamics applies history-dependent bias that disfavors previously visited states. To apply metadynamics, it necessary select few properties system─collective variables (CVs) can be used define potential. Over past years, there have been emerging opportunities for machine learning and, in particular, artificial neural networks within this domain. In broad context, specific unsupervised method was utilized, namely, parametric time-lagged t-distributed stochastic neighbor embedding (ptltSNE) design CVs. approach tested on Trp-cage trajectory (tryptophan cage) from literature. generate map conformations, distinguish fast conformational changes slow ones, and Then, metadynamic were performed. formation α-helix, we added α-RMSD collective variable. This simulation led one folding event 350 ns metadynamics simulation. degrees freedom not addressed CVs, performed parallel tempering metadynamics. 10 events 200 with 32 replicas.

Language: Английский

Citations

4
PiNN: Equivariant Neural Network Suite for Modeling Electrochemical Systems DOI Creative Commons
Jichen Li,

Lisanne Knijff,

Zhan‐Yun Zhang

et al.

Journal of Chemical Theory and Computation, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 30, 2025

Electrochemical energy storage and conversion play increasingly important roles in electrification sustainable development across the globe. A key challenge therein is to understand, control, design electrochemical materials with atomistic precision. This requires inputs from molecular modeling powered by machine learning (ML) techniques. In this work, we have upgraded our pairwise interaction neural network Python package PiNN via introducing equivariant features PiNet2 architecture for fitting potential surfaces along PiNet2-dipole dipole charge predictions as well PiNet2-χ generating atom-condensed response kernels. By benchmarking publicly accessible data sets of small molecules, crystalline materials, liquid electrolytes, found that shows significant improvements over original PiNet provides a state-of-the-art overall performance. Furthermore, leveraging on plug-ins such PiNNAcLe an adaptive learn-on-the-fly workflow ML potentials PiNNwall heterogeneous electrodes under external bias, expect serve versatile high-performing ML-accelerated platform systems.

Language: Английский

Citations

0
Machine learning of slow collective variables and enhanced sampling via spatial techniques DOI Open Access
Tuğçe Gökdemir, Jakub Rydzewski

Chemical Physics Reviews, Journal Year: 2025, Volume and Issue: 6(1)

Published: Feb. 3, 2025

Understanding the long-time dynamics of complex physical processes depends on our ability to recognize patterns. To simplify description these processes, we often introduce a set reaction coordinates, customarily referred as collective variables (CVs). The quality CVs heavily impacts comprehension dynamics, influencing estimates thermodynamics and kinetics from atomistic simulations. Consequently, identifying poses fundamental challenge in chemical physics. Recently, significant progress was made by leveraging predictive unsupervised machine learning techniques determine CVs. Many require temporal information learn slow that correspond long timescale behavior studied process. Here, however, specifically focus can identify corresponding slowest transitions between states without needing trajectories input, instead using spatial characteristics data. We discuss latest developments this category briefly potential directions for thermodynamics-informed

Language: Английский

Citations

0
Machine Learning of Slow Collective Variables and Enhanced Sampling via Spatial Techniques DOI Creative Commons
Tuğçe Gökdemir, Jakub Rydzewski

Published: Feb. 14, 2025

Understanding the long-time dynamics of complex physical processes depends on our ability to recognize patterns. To simplify description these processes, we often introduce a set reaction coordinates, customarily referred as collective variables (CVs). The quality CVs heavily impacts comprehension dynamics, influencing estimates thermodynamics and kinetics from atomistic simulations. Consequently, identifying poses fundamental challenge in chemical physics. Recently, significant progress was made by leveraging predictive unsupervised machine learning techniques determine CVs. Many require temporal information learn slow that correspond long timescale behavior studied process. Here, however, specifically focus can identify corresponding slowest transitions between states without needing trajectories input, instead using spatial characteristics data. We discuss latest developments this category briefly potential directions for thermodynamics-informed

Language: Английский

Citations

0
Learning Markovian dynamics with spectral maps DOI Open Access
Jakub Rydzewski, Tuğçe Gökdemir

The Journal of Chemical Physics, Journal Year: 2024, Volume and Issue: 160(9)

Published: March 4, 2024

The long-time behavior of many complex molecular systems can often be described by Markovian dynamics in a slow subspace spanned few reaction coordinates referred to as collective variables (CVs). However, determining CVs poses fundamental challenge chemical physics. Depending on intuition or trial and error construct lead non-Markovian with long memory effects, hindering analysis. To address this problem, we continue develop recently introduced deep-learning technique called spectral map [J. Rydzewski, J. Phys. Chem. Lett. 14, 5216-5220 (2023)]. Spectral learns maximizing gap Markov transition matrix describing anisotropic diffusion. Here, represent heterogeneous multiscale free-energy landscapes map, implement an adaptive algorithm estimate probabilities. Through state model analysis, validate that related the dominant relaxation timescales discerns between long-lived metastable states.

Language: Английский

Citations

3
Spectral Map for Slow Collective Variables, Markovian Dynamics, and Transition State Ensembles DOI Creative Commons
Jakub Rydzewski

Journal of Chemical Theory and Computation, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 12, 2024

Understanding the behavior of complex molecular systems is a fundamental problem in physical chemistry. To describe long-time dynamics such systems, which responsible for their most informative characteristics, we can identify few slow collective variables (CVs) while treating remaining fast as thermal noise. This enables us to simplify and treat it diffusion free-energy landscape spanned by CVs, effectively rendering Markovian. Our recent statistical learning technique, spectral map [Rydzewski, J.

Language: Английский

Citations

3
Dynamic framework for large-scale modeling of membranes and peripheral proteins DOI
Mohsen Sadeghi, David Rosenberger

Methods in enzymology on CD-ROM/Methods in enzymology, Journal Year: 2024, Volume and Issue: unknown, P. 457 - 514

Published: Jan. 1, 2024

Language: Английский

Citations

0
The algorithm for denoising point clouds of annular forgings based on Grassmann manifold and density clustering DOI
Yucun Zhang, An Wang,

Tao Kong

et al.

Measurement Science and Technology, Journal Year: 2024, Volume and Issue: 35(11), P. 115004 - 115004

Published: July 24, 2024

Abstract In the industrial sector, annular forgings serve as critical load-bearing components in mechanical equipment. During production process, precise measurement of dimensional parameters is paramount importance to ensure their quality and safety. However, owing influence environment, manufacturing process can introduce varying degrees noise, resulting inaccurate measurements. Therefore, researching methods for three-dimensional point cloud data eliminate noise forging clouds significant improving accuracy This paper presents a denoising approach based on Grassmann manifold density clustering (GDAD). First, within manifold, core points are determined using parameters. Second, performed with Cauchy distance replacing Euclidean reduce impact outliers analysis results. Finally, search tree model was constructed filter out incorrect clusters. The fusion results achieved data. Simulation experiments demonstrate that GDAD effectively eliminates edge performs well point-cloud models levels intensity.

Language: Английский

Citations

0