BioPathOpt

BioPathOpt is a Python package designed to streamline metabolic engineering workflows by integrating advanced modeling and optimization tools into a unified framework. It enables researchers to enrich stoichiometric metabolic models with additional biochemical data, seamlessly incorporate enzyme-constrained modeling approaches, and perform targeted optimization for the production of desired molecules.

Enzyme Constrained Modelling

Enzyme-constrained modelling extends traditional metabolic models by limiting reaction rates based on enzyme capacities. This adds biological realism, captures proteome allocation trade-offs, and improves prediction accuracy for metabolic engineering. There are two different modelling frameworks for this type of modelling GECKO written in MATLAB and ECMpy. The former is impractical if you do not have the right license and the later is slow and has not been updated for a while. To this end, we provide this package to help the metabolic engineering community run these types of models with more ease.

Below are plots that illustrate the improvements of this type of modelling framework to accurately simulate cell metabolism

Original iML1515 model predictions for various carbon sources	EC iML1515 model predictions for various carbon sources

We see that there are clear improvements:

Docker

The easiest way to use this project is to use the docker functionality that will open a notebook that enables you to run all aspects of this project. Below are three options:

Running with Docker Compose

This project ships with a flexible docker-compose.yml that lets you run BioPathOpt in three different ways:

1. Use a prebuilt image from Docker Hub
2. Build locally from images/Dockerfile.cache
3. Build locally from images/Dockerfile with a custom BRENDA_FILE argument

1. Run from Docker Hub (recommended)

Pulls the image directly from Docker Hub:

docker compose --profile hub up --build

2. Build from images/Dockerfile.cache

Uses the cached build recipe for faster local builds. This is because the repeated pubchem REST requests will make the image fail. We recommend that you use the following file to generate the cache before :

docker compose --profile cache up --build

3. Build from images/Dockerfile with BRENDA_FILE

Allows you to specify which Brenda flatfile to include at build time. Note that this will more likely fail because of the repeated REST requests to pubchem and we recommend that you either generate the cache locally, or use the image on dockerhub:

# Use default brenda_2024_1.json
docker compose --profile full up --build

# Or override with a custom file
BRENDA_FILE=brenda_2025_1.json docker compose --profile full up --build

Notes

• The JupyterLab server will be available at http://localhost:8085.
• Notebooks in your local ./notebooks/ directory are mounted inside the container at /home/notebooks/.
• No authentication token or password is set by default (--NotebookApp.token='' --NotebookApp.password='').

Installing

Run the following code to install by travelling to the github folder and running:

pip install -e .

Creating the cache

Important

You must manually download the brenda JSON on their website manually

The package parses the following databases: MetaNetX and Brenda and generates a large cache to enrich models. The later needs to be downloaded here and compressed:

from biopathopt.utils import write_compressed_json
import json
write_compressed_json(
    json.load(open('brenda_2024_1.json')), 
    'biopathopt/flatfiles/json_brenda.json.gz',
    )

To generate the cache run the following:

import biopathopt
data = biopathopt.Data()
data.refresh_cache()

Notebooks

Here is a non-exhaustive list of the features of this package.

Enrich model

The Read model notebook shows you how to pass any cobrapy model and enrich the annotations automatically

Enzyme constrained model

Enzyme-constrained notebook shows you how to build an enzyme constrained model and using an E.Coli model demonstrates the improvements in this modelling method vs the classical model by comparing it to actual measured data. (taken from ECMpy)

FSEOF Optimization

The FSEOF notebook (Flux Scanning based on Enforced Objective Flux) algorithm identifies metabolic reactions whose flux increases when production of a target compound is enforced, helping to pinpoint potential overexpression targets for strain engineering.

Acknowledgements

This package would not be possible without the work of ECMpy and FVSEOF where we I copied and altered a lot of the code. This is built relying heavily on CobraPy.

References

• Mao, Zhitao, et al. "ECMpy 2.0: A Python package for automated construction and analysis of enzyme-constrained models." Synthetic and systems biotechnology 9.3 (2024): 494-502.
• Ebrahim, Ali, et al. "COBRApy: constraints-based reconstruction and analysis for python." BMC systems biology 7.1 (2013): 74.
• Chen, Yu, et al. "Reconstruction, simulation and analysis of enzyme-constrained metabolic models using GECKO Toolbox 3.0." Nature protocols 19.3 (2024): 629-667.