Enzyme-constrained metabolic model of recombinant Pichia pastoris for hyaluronic acid production

This repository contains the computational workflow and data accompanying the manuscript:

Title: Recombinant HA Authors: Mehdi Sardari, Mohammadali bidari

The code implements an AutoPACMEN/sMOMENT-based workflow to build and calibrate an enzyme-constrained genome-scale model of recombinant Pichia pastoris expressing the hyaluronic-acid pathway, and to reproduce the simulations and figures reported in the manuscript.

Note about repository structure:
To keep the repository clean and reduce upload limitations on GitHub,
several large or automatically generated files (e.g., AutoPACMEN spreadsheets, intermediate JSON files, and some simulation results)
You can extract the ZIP file after cloning the repository.
The internal folder structure of the ZIP matches the layout described below (code/, model/, data/, results/).

Repository structure

Recombinant_HA/
├── code/
│ ├── adding annotations.ipynb
│ ├── Create Ec model.ipynb
│ ├── Calibration.ipynb
│ ├── Validation.ipynb
│ └── results.ipynb
│ ├── data/
│ ├── bigg_models_metabolites.txt
│ └── bigg_models_reactions.txt
│ ├── model/
│ ├── model_recombinant_ready_for_smoment.xml
│ ├── model_annotated.xml
│ └── model_annotated_ec_05_10_09.xml
│ └── results/
├── (simulation figures)
└── AutoPacmen Output/
├── A.json
├── bigg_id_name_mapping.json
├── brenda_kcat.json
├── combined_kcat_model.json
├── MM_compartments.xlsx
├── MM_enzyme_stoichiometries.xlsx
├── MM_metabolite_concentrations.xlsx
├── MM_metabolites.xlsx
├── MM_protein_data.xlsx
├── MM_reactions.xlsx
├── MM_reactions_kcat_mapping_combined.json
└── sabio_model.json

Notebooks Overview (Workflow)

code/adding annotations.ipynb

Adds EC numbers, UniProt IDs, and other annotations to the recombinant P. pastoris model.
Produces:

• model/model_annotated.xml

code/Create Ec model.ipynb

Generates enzyme-constrained model using AutoPACMEN.
Outputs include:

• MM spreadsheets in results/AutoPacmen Output/
• model/model_annotated_ec_05_10_09.xml (Final ecModel after Check in Calibration and Validation Step)

code/Calibration.ipynb

Calibrates the enzyme-constrained model using experimental biomass and CO₂ production data.
Figures are saved in results/.

code/Validation.ipynb

Validates HA production and growth rate against literature values.
Produces:

• results/Hyaluronic Acid Production - HA Validation.png

code/results.ipynb

Reproduces all simulations in the manuscript:

• Carbon source effects
• Amino acid groups
• Nitrogen sources
• Methanol–glucose co-feeding
• Pareto front between HA and biomass

All figures are saved in results/.

Models

model/model_recombinant_ready_for_smoment.xml

Base recombinant model including the HA pathway.

model/model_annotated.xml

Model after adding structural and functional annotations.

model/model_annotated_ec_05_10_09.xml

Final enzyme-constrained sMOMENT model after calibration used in validation, and simulations.

External Databases & Missing Large Files

Note about external databases:
The enzyme kinetic data used in this project were originally retrieved from the
BRENDA 2025 database.
Due to size and licensing restrictions, the raw BRENDA dataset and the full extracted JSON file
(generated during the AutoPACMEN preprocessing stage) are not included in this repository.

Instead, the repository contains the processed and minimal JSON/XLSX files inside
results/AutoPacmen Output/, which are sufficient to rebuild the enzyme-constrained model
(model_annotated_ec_05_10_09.xml) and to run all simulations in the manuscript.

Users who wish to fully reproduce the AutoPACMEN pipeline from scratch may download
the corresponding BRENDA release manually from:
https://www.brenda-enzymes.org
(registration required), and place the extracted files in the project directory as instructed in the
Create Ec model.ipynb notebook.

Requirements

The workflow was tested using:

• Python 3.10
• cobra (COBRApy)
• numpy
• pandas
• matplotlib
• AutoPACMEN (latest GitHub version)