This project aims to predict metabolic fluxes from omics data with machine learning models.
• /plots - This directory contains all graphical results, including figures that are referenced in the published article
• /results - This directory contains a detailed set of results from all predictors, including predicted flux values and statistical significance test results against pFBA
• /sbml - SBML model(s) used for pFBA simulations are stored in this folder
• analysis.py
- This script is responsible for the analytical post-processing of the results
• data.py
- Contains data sources and relevant configurations for data loading
• learning.py
- This script encapsulates the logic for implementing Neural Networks
• models.py
- This script performs all tasks related to model fitting, cross-validation, and independent testing
• pfba.py
- Script that performs pFBA simulations
• plots.py
- Generates all the figures found in the /plots directory
• requirements.txt - Lists the packages required to replicate the development environment, e.g., using conda
• utils.py
- Contains miscellaneous auxiliary functions.
This archive contains results, code and data utilized in the paper: "Predicting metabolic fluxes from omics data via machine learning: Moving from knowledge-driven towards data-driven approaches". Computational and Structural Biotechnology Journal (2023) | DOI: https://doi.org/10.1016/j.csbj.2023.10.002