Code and figures for "Modelling dynamic host-pathway interactions at the genome scale with machine learning"

Modelling dynamic host-pathway interactions at the genome scale This repository is for the joint FBA-ODE simulator method described in the recent paper "Modelling dynamic host-pathway interactions at the genome scale". Data files are too large for GitHub and available on request. To cite this work, please reference: Modelling dynamic host-pathway interactions at the genome scale. by Charlotte Merzbacher, Oisin Mac Aodha, and Diego Oyarzún. Metabolic Engineering (2025). Requirements This code is written primarily in Julia 1.8.x and uses the following packages: DifferentialEquations COBREXA DataFrames Tulip Plots Colors ModelingToolkit Statistics GLM Random Flux ProgressMeter MLBase #Confusion matrix function Serialization TreeParzen CSV LatinHypercubeSampling Visualization code is written in Python 3.x and uses the following packages: pandas matplotlib seaborn numpy colormaps os statistics Model files models/beta_carotene.jl Julia implementation of beta-carotene ODE model. models/glucaric_acid.jl Julia implementation of glucaric acid ODE model. models/iML1515.xml SBML model of iML1515 GSM, downloaded from BIGG models database. models/ml_models/ga/ feas_model.jls Logistic regression model for prediction of FBA feasibility for glucaric acid lam_model.jls Linear regression model for prediction of growth rate for glucaric acid v_in_model.jls Neural network model for prediction of boundary flux for glucaric acid models/ml_models/bcar/ feas_model.jls Logistic regression model for prediction of FBA feasibility for beta-carotene lam_model.jls Linear regression model for prediction of growth rate for beta-carotene v_in_model.jls Linear regressionmodel for prediction of boundary flux component 1 (influx to IPP) for beta-carotene v_ipp_model.jls Linear regression model for prediction of boundary flux component 2 (efflux from IPP) for beta-carotene v_fpp_model.jls Linear regression model for prediction of boundary flux component 3 (efflux from FPP) for beta-carotene Experiment code experiments/bcar_experiments.jl Julia code to run all experiments with beta-carotene model. experiments/ga_experiments.jl Julia code to run all experiments with glucaric model. Note that filepaths are at the top of the files and must be changed when repo is cloned to allow code to find necessary data files. Functions expect appropriate folders have already been created with correct names. All functions have docstrings which give information about what they do and the necessary inputs (if any). Visualization notebooks timing study/timing_study.ipynb Simulator runtime experiment (in Julia) and visualization (in Julia) of results for supplementary figures 1 and 2 visualization/bcar_visualization.ipynb Visualization code (in Python) of all results from beta-carotene case study. Includes code to generate figures 2d, 3c, d, e, f, and 4b, c. visualization/ga_visualization.ipynb Visualization code (in Python) of all results from glucaric acid case study. Includes code to generate figures 2b, 3b, and 4d

Merzbacher, C. (2025). Code and figures for "Modelling dynamic host-pathway interactions at the genome scale with machine learning". Zenodo. https://doi.org/10.5281/zenodo.15518931