We introduce a Bayesian framework leveraging synthetic likelihoods to enable uncertainty quantification and robust inference of non-bonded force parameters in three-point water models.The approach integrates multiple ...We introduce a Bayesian framework leveraging synthetic likelihoods to enable uncertainty quantification and robust inference of non-bonded force parameters in three-point water models.The approach integrates multiple experimental observables—enthalpy of vaporization,molecular volume,the radial distribution function,and hydrogen bonding patterns—to explicitly infer model parameters.Beyond parameter estimation,we quantify uncertainty in both inference observables and validation properties,including those that are difficult to target by other means.By systematically analyzing the response of these observables to parameter variations,our method highlights inherent limitations of three-point water models.These findings highlight the utility of our framework in integrating diverse data sources in a principled uncertainty quantification workflow,ultimately improving confidence in the ability of molecular dynamics simulations to reproduce experimental data.Additionally,we evaluate the performance of the mean and the mode of the posterior distribution,demonstrating the limitations of this family of models.展开更多
Structural assumptions in infectious disease models,such as the choice of network or compartmental model type or the inclusion of different types of heterogeneity across individuals,might affect model predictions as m...Structural assumptions in infectious disease models,such as the choice of network or compartmental model type or the inclusion of different types of heterogeneity across individuals,might affect model predictions as much as or more than the choice of input parameters.We explore the potential implications of structural assumptions on HIV model predictions and policy conclusions.We illustrate the value of inference robustness assessment through a case study of the effects of a hypothetical HIV vaccine in multiple population subgroups over eight related transmission models,which we sequentially modify to vary over two dimensions:parameter complexity(e.g.,the inclusion of age and HCV comorbidity)and contact/simulation complexity(e.g.,aggregated compartmental vs.individual/disaggregated compartmental vs.network models).We find that estimates of HIV incidence reductions from network models and individual compartmental models vary,but those differences are overwhelmed by the differences in HIV incidence between such models and the aggregated compartmental models(which aggregate groups of individuals into compartments).Complexities such as age structure appear to buffer the effects of aggregation and increase the threshold of net vaccine effectiveness at which aggregated models begin to overestimate reductions.The differences introduced by parameter complexity in estimated incidence reduction also translate into substantial differences in cost-effectiveness estimates.Parameter complexity does not appear to play a consistent role in differentiating the projections of network models.展开更多
基金This research was supported financially by the project AI4Research at Uppsala University,Sweden,and by the Swedish Research Council(grants 2020-05059,2024-04314)Funding from eSSENCE-The e-Science Collaboration(Uppsala-Lund-Ume˚a,Sweden)is gratefully acknowledged.Additional funding and support were provided by the Centre for Interdisciplinary Mathematics(CIM)at Uppsala UniversityComputer resources provided by the National Academic Infrastructure for Supercomputing Sweden at the PDC Center for High Performance Computing,KTH Royal Institute of Technology,Sweden,partially funded by the Swedish Research Council through(grant 2022-06725).We also acknowledge the Molecular Biophysics program at Uppsala University for providing access to local computational resources.
文摘We introduce a Bayesian framework leveraging synthetic likelihoods to enable uncertainty quantification and robust inference of non-bonded force parameters in three-point water models.The approach integrates multiple experimental observables—enthalpy of vaporization,molecular volume,the radial distribution function,and hydrogen bonding patterns—to explicitly infer model parameters.Beyond parameter estimation,we quantify uncertainty in both inference observables and validation properties,including those that are difficult to target by other means.By systematically analyzing the response of these observables to parameter variations,our method highlights inherent limitations of three-point water models.These findings highlight the utility of our framework in integrating diverse data sources in a principled uncertainty quantification workflow,ultimately improving confidence in the ability of molecular dynamics simulations to reproduce experimental data.Additionally,we evaluate the performance of the mean and the mode of the posterior distribution,demonstrating the limitations of this family of models.
基金Financial support for this study was provided by Grant Number R01-DA15612 from the National Institute on Drug AbuseCB was supported by a PACCAR Inc.Stanford Graduate Fellowship and National Science Foundation Graduate Fellowship DGE-114747.
文摘Structural assumptions in infectious disease models,such as the choice of network or compartmental model type or the inclusion of different types of heterogeneity across individuals,might affect model predictions as much as or more than the choice of input parameters.We explore the potential implications of structural assumptions on HIV model predictions and policy conclusions.We illustrate the value of inference robustness assessment through a case study of the effects of a hypothetical HIV vaccine in multiple population subgroups over eight related transmission models,which we sequentially modify to vary over two dimensions:parameter complexity(e.g.,the inclusion of age and HCV comorbidity)and contact/simulation complexity(e.g.,aggregated compartmental vs.individual/disaggregated compartmental vs.network models).We find that estimates of HIV incidence reductions from network models and individual compartmental models vary,but those differences are overwhelmed by the differences in HIV incidence between such models and the aggregated compartmental models(which aggregate groups of individuals into compartments).Complexities such as age structure appear to buffer the effects of aggregation and increase the threshold of net vaccine effectiveness at which aggregated models begin to overestimate reductions.The differences introduced by parameter complexity in estimated incidence reduction also translate into substantial differences in cost-effectiveness estimates.Parameter complexity does not appear to play a consistent role in differentiating the projections of network models.
