🧬 Baysian Inference for Biological ODEs¶
Bayesian Learning of Microbial Traits from Population Time Series Data: A Primer¶
Mathematical models are increasingly used to infer traits, interactions, and functional dynamics of microbial systems. The inference process typically begins with a rate-based Ordinary Differential Equation (ODE) model.
However, fitting such models to experimental data requires a principled statistical framework that can:
Incorporate prior knowledge, Account for measurement noise, and Quantify uncertainty in parameter estimates.
Our aim
We strive to make the implicit, explicit — introducing Bayesian inference of ecological ODE models for microbial time series, with a unified workflow in Python (PyMC) and Julia (Turing.jl).
⬇️ Download the full repository¶
Get all notebooks, code, and examples in one click:
📘 Overview¶
This repository accompanies our upcoming paper:
“Bayesian Learning of Microbial Traits from Population Time Series Data: A Primer”
Authors: Raunak Dey, Robert Beach, Kennedi M. Hambrick, Ioannis Sgouralis, Paul Fremont, David Demory, Eric Carr, Stephen J. Beckett, Joshua S. Weitz, David Talmy (Link will be posted here when the paper is online.)
🚀 Case studies¶
Python (PyMC)¶
You can view the solved examples (case studies) or try running it yourself via Google Colab.
-
Case Study 1 — Exponential Growth and Death
📄 View notebook on GitHub
🧑💻 Try it yourself:
-
Case Study 2 — Logistic Growth and Death
📄 View notebook on GitHub
🧑💻 Try it yourself: -
Case Study 3 — Monod Growth and Death
📄 View notebook on GitHub
🧑💻 Try it yourself:
Julia (Turing.jl)¶
- Case Study 1 — Exponential Growth and Death
- Case Study 2 — Logistic Growth and Death
- Case Study 3 — Monod Growth and Death
🧰 Stack¶
- Python · PyMC — Probabilistic programming in Python for Bayesian modeling and inference
- Julia · Turing.jl — A flexible probabilistic programming language in Julia
All analyses were performed using:
Python
- Python 3.12
- PyMC 5.25.1
- PyTensor 2.31.7
- ArviZ 0.22.0
- NumPy 2.2.5
- SciPy 1.16.2
Julia
- Julia 1.11.0
- Turing.jl 0.40.2
- DifferentialEquations.jl 7.16.1
- SciMLSensitivity.jl 7.90.0
- Distributions.jl 0.25.120
- MCMCChains.jl 7.2.0
👩🔬 Contributors¶
This primer was developed through a collaborative effort across multiple research groups, combining expertise in microbial ecology, statistical physics, and Bayesian inference.
Raunak Dey — University of Maryland
🔗 Website · GitHub
David Talmy — University of Tennessee, Knoxville
🔗 Website
Robert Beach — University of Tennessee, Knoxville
Kennedi Hambrick — University of Tennessee, Knoxville
Ioannis Sgouralis — University of Tennessee, Knoxville
🔗 Website
Stephen J. Beckett — University of Maryland 🔗 Website
Paul Frémont — University of Maryland 🔗 Website
David Demory — Sorbonne Université 🔗 Website
Eric Carr — University of Tennessee, Knoxville
Joshua S. Weitz — University of Maryland 🔗 Website
This project connects theory, data, and computation to advance reproducible Bayesian inference for ecological population models.