Bayesian-Inference-Driven Model Parametrization and Model Selection for 2CLJQ Fluid Models

Abstract

A high level of physical detail in a molecular model improves its ability to perform high accuracy simulations but can also significantly affect its complexity and computational cost. In some situations, it is worthwhile to add complexity to a model to capture properties of interest; in others, additional complexity is unnecessary and can make simulations computationally infeasible. In this work, we demonstrate the use of Bayesian inference for molecular model selection, using Monte Carlo sampling techniques accelerated with surrogate modeling to evaluate the Bayes factor evidence for different levels of complexity in the two-centered Lennard-Jones + quadrupole (2CLJQ) fluid model. Examining three nested levels of model complexity, we demonstrate that the use of variable quadrupole and bond length parameters in this model framework is justified only for some chemistries. Through this process, we also get detailed information about the distributions and correlation of parameter values, enabling improved parametrization and parameter analysis. We also show how the choice of parameter priors, which encode previous model knowledge, can have substantial effects on the selection of models, penalizing careless introduction of additional complexity. We detail the computational techniques used in this analysis, providing a roadmap for future applications of molecular model selection via Bayesian inference and surrogate modeling.

Machine Learning Directed Optimization of Classical Molecular Modeling Force Fields

“…This naturally raises another question: given different atom-typing schemes, which should be used? Recent work demonstrates the promise of using Bayes factors to compare models with different levels of complexity (e.g., different atom-typing schemes) and make a justified selection.…”

“…This naturally raises another question: given different atom-typing schemes, which should be used? Recent work 90 demonstrates the promise of using Bayes factors to compare models with different levels of complexity (e.g., different atom-typing schemes) and make a justified selection. Since the prior analysis was performed entirely with the predictions of the GP surrogate models, we performed molecular simulations with two top-performing parameter sets for each of the shared atom-typing schemes (AT-2, AT-3, and AT-4) in order to compute the simulated MAPE values and compare them with the surrogate model predictions.…”

TIES 2.0: A Dual-Topology Open Source Relative Binding Free Energy Builder with Web Portal

“…Further automation of the process will increase the potential user base, an important factor in the ever more expensive search for novel medicines. As we witness further improvements in the workflows, , automated and systematic forcefield development, computational costs, targets and their variations, , together with accessibility and automation, we expect to encounter more real-life applications and successes.…”

End-to-end differentiable construction of molecular mechanics force fields

“…While most attention has been focused on the continuous parameters of the force field model with fixed model form, some progress has been made in discrete model selection among candidate model forms. 136–138…”