sbi: A toolkit for simulation-based inference

Tejero-Cantero, Álvaro; Boelts, Jan; Deistler, Michael; Lueckmann, Jan-Matthis; Durkan, Conor; Gonçalves, Pedro J.; Greenberg, David S.; Macke, Jakob H.

doi:10.21105/joss.02505

Cited by 165 publications

(136 citation statements)

References 11 publications

Supporting

Mentioning

136

Contrasting

Order By: Relevance

“…Code implementing SNPE based on Theano, is available at http://www.mackelab.org/delfi/ . An extended toolbox based on PyTorch is available at http://www.mackelab.org/sbi/ ( Tejero-Cantero et al, 2020 ).…”

Section: Methodsmentioning

confidence: 99%

Training deep neural density estimators to identify mechanistic models of neural dynamics

Gonçalves

Lueckmann

Deistler

et al. 2020

eLife

Self Cite

169

220

View full text Add to dashboard Cite

Mechanistic modeling in neuroscience aims to explain observed phenomena in terms of underlying causes. However, determining which model parameters agree with complex and stochastic neural data presents a significant challenge. We address this challenge with a machine learning tool which uses deep neural density estimators- trained using model simulations- to carry out Bayesian inference and retrieve the full space of parameters compatible with raw data or selected data features. Our method is scalable in parameters and data features, and can rapidly analyze new data after initial training. We demonstrate the power and flexibility of our approach on receptive fields, ion channels, and Hodgkin-Huxley models. We also characterize the space of circuit configurations giving rise to rhythmic activity in the crustacean stomatogastric ganglion, and use these results to derive hypotheses for underlying compensation mechanisms. Our approach will help close the gap between data-driven and theory-driven models of neural dynamics.

show abstract

Section: Methodsmentioning

confidence: 99%

Training deep neural density estimators to identify mechanistic models of neural dynamics

Gonçalves

Lueckmann

Deistler

et al. 2020

eLife

Self Cite

169

220

View full text Add to dashboard Cite

show abstract

“…We applied NPE ( Figure 2B ), implemented in the Python package sbi [81], and tested two specialized normalizing flows as density estimators: a masked autoregressive flow (MAF) [93] and a neural spline flow (NSF) [94] . The normalizing flow is used as a density estimator to “learn” an amortized posterior distribution, which can then be evaluated for specific observations.…”

Section: Resultsmentioning

confidence: 99%

“…When directly inferring the DFE, we used a uniform prior for α from 0.5 to 15 and a log-uniform prior for β from 10 −3 to 0.8. We held out three experiments from the set of eleven and used a three-layer neural network to reduce the remaining eight observations to a five-feature summary statistic vector, which we then used as an embedding net [81] with NPE to infer the joint posterior distribution of α, β , and δ C ( Figure 5 , blue lines). For each observation set, we performed each inference method three times, using different sets of eight experiments to infer the underlying DFE.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Simulation-based inference of evolutionary parameters from adaptation dynamics using neural networks

Avecilla

Gresham

Ram

2021

Preprint

View full text Add to dashboard Cite

The rate of adaptive evolution depends on the rate at which beneficial mutations are introduced into a population and the fitness effects of those mutations. The rate of beneficial mutations and their expected fitness effects is often difficult to empirically quantify. As these two parameters determine the pace of evolutionary change in a population, the dynamics of adaptive evolution may enable inference of their values. Copy number variants (CNVs) are a pervasive source of heritable variation that can facilitate rapid adaptive evolution. Previously, we developed a locus-specific fluorescent CNV reporter to quantify CNV dynamics in evolving populations maintained in nutrient-limiting conditions using chemostats. Here, we use the observed CNV adaptation dynamics to estimate the rate at which beneficial CNVs are introduced through de novo mutation and their fitness effects using simulation-based Bayesian likelihood-free inference approaches. We tested the suitability of two evolutionary models: a standard Wright-Fisher model and a chemostat growth model. We evaluated two likelihood-free inference algorithms: the well-established Approximate Bayesian Computation with Sequential Monte Carlo (ABC-SMC) algorithm, and the recently developed Neural Posterior Estimation (NPE) algorithm, which applies an artificial neural network to directly estimate the posterior distribution. By systematically evaluating the suitability of different inference methods and models we show that NPE has several advantages over ABC-SMC and that a Wright-Fisher evolutionary model suffices in most cases. Using our validated inference framework, we estimate the CNV formation rate at the GAP1 locus in yeast as 10-4.7 -10-4 per cell division, and a selection coefficient of 0.04 - 0.1 per generation for GAP1 CNVs in glutamine-limited chemostats. We experimentally validated our estimates using barcode lineage tracking and pairwise fitness assays. Our results are consistent with a high beneficial CNV supply rate that is 10-fold greater than the estimated rates of beneficial single-nucleotide mutations, explaining their outsized importance in rapid adaptive evolution. More generally, our study demonstrates the utility of novel simulation-based likelihood-free inference methods for inferring the rates and effects of evolutionary processes from empirical data.

show abstract

“…These approaches are computationally heavy, often requiring a simulator and extensive computer codes. Open-access packages for such simulation-based approaches makes these methods more accessible ( Tejero-Cantero et al., 2020 ). Here, we describe the use of a simulator of synaptic electrophysiology experiments along with a calculation of synthetic likelihoods used for the estimation of silent synapse fraction.…”

Section: Before You Beginmentioning

confidence: 99%

Accurate Silent Synapse Estimation from Simulator-Corrected Electrophysiological Data Using the SilentMLE Python Package

2020

View full text Add to dashboard Cite

Summary The proportion of silent (AMPAR-lacking) synapses is thought to be related to the plasticity potential of neural networks. We created a maximum-likelihood estimator of silent synapse fraction based on simulations of the underlying experimental methodology. Here, we provide a set of guidelines for running a Python package on compatible experimental synaptic data. Compared with traditional failure-rate approaches, this synthetic likelihood estimator improves the validity and accuracy of the estimates of the silent synapse fraction. For complete details on the use and execution of this protocol, please refer to Lynn et al. (2020) .

show abstract

sbi: A toolkit for simulation-based inference

Cited by 165 publications

References 11 publications

Training deep neural density estimators to identify mechanistic models of neural dynamics

Training deep neural density estimators to identify mechanistic models of neural dynamics

Simulation-based inference of evolutionary parameters from adaptation dynamics using neural networks

Accurate Silent Synapse Estimation from Simulator-Corrected Electrophysiological Data Using the SilentMLE Python Package

Contact Info

Product

Resources

About