Multiscale Reweighted Stochastic Embedding: Deep Learning of Collective Variables for Enhanced Sampling

@article{Rydzewski2021MultiscaleRS,
  title={Multiscale Reweighted Stochastic Embedding: Deep Learning of Collective Variables for Enhanced Sampling},
  author={Jakub Rydzewski and Omar Valsson},
  journal={The Journal of Physical Chemistry. a},
  year={2021},
  volume={125},
  pages={6286 - 6302}
}
Machine learning methods provide a general framework for automatically finding and representing the essential characteristics of simulation data. This task is particularly crucial in enhanced sampling simulations. There we seek a few generalized degrees of freedom, referred to as collective variables (CVs), to represent and drive the sampling of the free energy landscape. In theory, these CVs should separate different metastable states and correspond to the slow degrees of freedom of the… 
Discover, Sample, and Refine: Exploring Chemistry with Enhanced Sampling Techniques.
Over the last few decades, enhanced sampling methods have been continuously improved. Here, we exploit this progress and propose a modular workflow for blind reaction discovery and determination of
MLCV: Bridging Machine-Learning-Based Dimensionality Reduction and Free-Energy Calculation
TLDR
A user-friendly tool that facilitates the use of machine-learned CVs in importance-sampling simulations through the popular Colvars module, and believes that MLCV provides an effective, powerful, and user- friendly platform accessible to experts and nonexperts alike for machine-learning (ML)-guided CV discovery and enhanced-sampled simulations to unveil the molecular mechanisms underlying complex biochemical processes.
Discover, Sample and Refine: Exploring Chemistry with Enhanced Sampling Techniques
Over the last few decades enhanced sampling methods have made great strides. Here, we exploit this progress and propose a modular workflow for blind reaction discovery and characterization of
Deep learning the slow modes for rare events sampling
TLDR
The power of machine learning and the efficiency of the recently developed on the fly probability-enhanced sampling method are brought to bear on this approach, resulting in a powerful and robust algorithm that removes the bottlenecks that hinder convergence.
SGOOP-d: Estimating Kinetic Distances and Reaction Coordinate Dimensionality for Rare Event Systems from Biased/Unbiased Simulations.
TLDR
This work develops a formalism that learns a multidimensional yet minimally complex reaction coordinate (RC) for generic high-dimensional systems and shows how this method can work on long unbiased simulations as well as biased simulations often needed for rare event systems.
Achieving 100X faster simulations of complex biological phenomena by coupling ML to HPC ensembles
TLDR
DeepDriveMD is established as a high-performance framework for ML-driven HPC simulation scenarios, that supports diverse simulation and ML back-ends, and is extended to support stream-based communication between simulations and learning methods.
Coupling streaming AI and HPC ensembles to achieve 100-1000x faster biomolecular simulations
TLDR
The results establish DeepDriveMD as a high-performance framework for ML-driven HPC simulation scenarios, that supports diverse MD simulation and ML back-ends, and which enables new scientific insights by improving the length and time scales accessible with current computing capacity.

References

SHOWING 1-10 OF 166 REFERENCES
Physics-Inspired Structural Representations for Molecules and Materials.
TLDR
This review summarizes the current understanding of the nature and characteristics of the most commonly used structural and chemical descriptions of atomistic structures, highlighting the deep underlying connections between different frameworks and the ideas that lead to computationally efficient and universally applicable models.
Improving sample and feature selection with principal covariates regression
TLDR
Two popular subselection schemes applied to this end, CUR decomposition, derived from a low-rank approximation of the feature matrix, and farthest point sampling, are modified, incorporating a supervised component following the same spirit as the principal covariates (PCov) regression method.
Global Free-Energy Landscapes as a Smoothly Joined Collection of Local Maps.
TLDR
This paper presents the Adaptive Topography of Landscape for Accelerated Sampling (ATLAS), a new biasing method capable of working with many CVs, to apply a divide-and-conquer strategy.
A statistical physics based approach to capture spurious solutions in AI-enhanced molecular dynamics
TLDR
A new and automated algorithm is reported based on the notion that a more reliable AI solution for many problems in chemistry and biophysics will be one that maximizes the time scale separation between slow and fast processes, which will lead to increased and robust use of trustworthy AI in molecular simulations of complex systems.
Perplexity-free Parametric t-SNE
TLDR
A multi-scale parametric t-SNE scheme, relieved from the perplexity tuning and with a deep neural network implementing the mapping, competitive with the best perplexity adjustments in terms of neighborhood preservation on multiple data sets.
Identification of ligand-specific G protein-coupled receptor states and prediction of downstream efficacy via data-driven modeling
TLDR
The results demonstrate that computational methods can be used to assess GPCR drug candidates with specific efficacy profiles and find an excellent correlation between the conformation of the microswitches close to the ligand binding site and in the transmembrane region and experimentally reported cAMP signaling responses, highlighting the predictive power of the apprch.
Unified Approach to Enhanced Sampling
TLDR
This work introduces a new class of collective-variables-based bias potentials that can be used to sample any of the expanded ensembles normally sampled via replica exchange, and provides a practical implementation of the recently developed on-the-fly probability enhanced sampling method.
Confronting pitfalls of AI-augmented molecular dynamics using statistical physics
TLDR
A novel and automated algorithm based on the notion that a more reliable AI solution for many problems in chemistry and biophysics will be one that maximizes the time scale separation between slow and fast processes is reported, based on a maximum path entropy or caliber-based framework.
Machine learning force fields and coarse-grained variables in molecular dynamics: application to materials and biological systems.
TLDR
A review of the current understanding of goals, benefits, and limitations of machine learning techniques for computational studies on atomistic systems, focusing on the construction of empirical force fields from ab-initio databases and the determination of reaction coordinates for free energy computation and enhanced sampling.
Data Reweighting in Metadynamics Simulations.
TLDR
It is shown that the free energy landscape reconstructed from reweighted data can be more accurate than the negative bias potential depending on the reweighting technique, the stage of the simulation and the adoption of well-tempered or standard metadynamics.
...
...