GpABC: a Julia package for approximate Bayesian computation with Gaussian process emulation

Tankhilevich, Evgeny; Ish-Horowicz, Jonathan; Hameed, Tara; Roesch, Elisabeth; Kleijn, Istvan T.; Stumpf, Michael; He, Fei

doi:10.1093/bioinformatics/btaa078

Cited by 21 publications

(20 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…implemented in Stan ( Carpenter et al , 2017 ) and BCM ( Thijssen et al , 2016 )], or e.g. ABC methods exploiting Gaussian processes ( Tankhilevich et al , 2020 ). In general, alternative methods may be limited in their applicability, e.g.…”

Section: Discussionmentioning

confidence: 99%

Efficient exact inference for dynamical systems with noisy measurements using sequential approximate Bayesian computation

Schälte

Hasenauer

2020

Bioinformatics

View full text Add to dashboard Cite

Motivation Approximate Bayesian computation (ABC) is an increasingly popular method for likelihood-free parameter inference in systems biology and other fields of research, as it allows analyzing complex stochastic models. However, the introduced approximation error is often not clear. It has been shown that ABC actually gives exact inference under the implicit assumption of a measurement noise model. Noise being common in biological systems, it is intriguing to exploit this insight. But this is difficult in practice, as ABC is in general highly computationally demanding. Thus, the question we want to answer here is how to efficiently account for measurement noise in ABC. Results We illustrate exemplarily how ABC yields erroneous parameter estimates when neglecting measurement noise. Then, we discuss practical ways of correctly including the measurement noise in the analysis. We present an efficient adaptive sequential importance sampling-based algorithm applicable to various model types and noise models. We test and compare it on several models, including ordinary and stochastic differential equations, Markov jump processes and stochastically interacting agents, and noise models including normal, Laplace and Poisson noise. We conclude that the proposed algorithm could improve the accuracy of parameter estimates for a broad spectrum of applications. Availability and implementation The developed algorithms are made publicly available as part of the open-source python toolbox pyABC (https://github.com/icb-dcm/pyabc). Supplementary information Supplementary data are available at Bioinformatics online.

show abstract

Section: Discussionmentioning

confidence: 99%

Efficient exact inference for dynamical systems with noisy measurements using sequential approximate Bayesian computation

Schälte

Hasenauer

2020

Bioinformatics

View full text Add to dashboard Cite

show abstract

“…Although it is very convenient for us to utilise the likelihood function based on our time-dependent solution, other likelihood-free calibration methods are available. One of particular mention is approximate Bayesian computation (ABC), which generates SSA (Monte Carlo) trajectories for a given parameter set θ which can then be compared to the original data using a distance function [59,60]. The choice of distance function is important, and typical examples include the Euclidean distance between the trajectories, the Hellinger distance, or even the Wasserstein distance [61].…”

Section: Other Methods For Model Calibrationmentioning

confidence: 99%

“…In our case, even with an analytically defined likelihood function, and only three parameters to infer, model calibration is still a difficult task unless one utilises a wellinformed data-set. Additionally, as we have already commented, it is possible to conduct calibration without an explicit likelihood function using likelihood-free methods for parameter inference [60]. It is hence vital for model calibration that either calibration methods become much smarter, for example using methods beyond simply using a likelihood function as done in [67] (here in the context of gene expression), or else the data used for calibration becomes more informative.…”

Section: Implications Of Calibration Proceduresmentioning

confidence: 99%

Non-equilibrium time-dependent solution to discrete choice with social interactions

Holehouse,

Pollitt

2021

Preprint

View full text Add to dashboard Cite

We solve the binary decision model of Brock and Durlauf [1] in time using methods often employed in studies of stochastic complex systems. This solution is valid when not at equilibrium and can be used to exemplify path-dependent behaviours of their model. The solution is computationally fast and is indistinguishable from Monte Carlo simulation. Lock-in effects are observed in some regions of the model's parameter space, and we calculate the time scale of the lock-ins. Curiously, we find that although altruistic agents coalesce more strongly on a particular decision, increasing their utility in the short-term, they are also more prone to being stuck in a non-optimal decision lock-in as compared to selfish agents. Finally, we construct a likelihood function that can be used on non-equilibrium data for model calibration. Even with a well-defined likelihood function, model calibration is difficult unless one has access to data representative of the underlying model.

show abstract

“…As in previous sections, we examine x = {0.1, 1, 5, 50}. For each x , we infer for y = {0.3, 1, 1.7} using the GpABC Julia package [ 47 ]. Accepted x , y values from the final ABC SMC population make up the posterior distributions for x and y , with the means taken to be the inferred parameter values.…”

Section: Model Comparison and Parameter Inferencementioning

confidence: 99%

Cluster mean-field theory accurately predicts statistical properties of large-scale DNA methylation patterns

Kerr

Sproul

Grima

2022

J. R. Soc. Interface.

View full text Add to dashboard Cite

The accurate establishment and maintenance of DNA methylation patterns is vital for mammalian development and disruption to these processes causes human disease. Our understanding of DNA methylation mechanisms has been facilitated by mathematical modelling, particularly stochastic simulations. Megabase-scale variation in DNA methylation patterns is observed in development, cancer and ageing and the mechanisms generating these patterns are little understood. However, the computational cost of stochastic simulations prevents them from modelling such large genomic regions. Here, we test the utility of three different mean-field models to predict summary statistics associated with large-scale DNA methylation patterns. By comparison to stochastic simulations, we show that a cluster mean-field model accurately predicts the statistical properties of steady-state DNA methylation patterns, including the mean and variance of methylation levels calculated across a system of CpG sites, as well as the covariance and correlation of methylation levels between neighbouring sites. We also demonstrate that a cluster mean-field model can be used within an approximate Bayesian computation framework to accurately infer model parameters from data. As mean-field models can be solved numerically in a few seconds, our work demonstrates their utility for understanding the processes underpinning large-scale DNA methylation patterns.

show abstract

GpABC: a Julia package for approximate Bayesian computation with Gaussian process emulation

Cited by 21 publications

References 4 publications

Efficient exact inference for dynamical systems with noisy measurements using sequential approximate Bayesian computation

Efficient exact inference for dynamical systems with noisy measurements using sequential approximate Bayesian computation

Non-equilibrium time-dependent solution to discrete choice with social interactions

Cluster mean-field theory accurately predicts statistical properties of large-scale DNA methylation patterns

Contact Info

Product

Resources

About