This repository contains code to reproduce simulations related to enhancing the accuracy of genome-scale metabolic models by incorporating kinetic information. We started from a genome-scale and a kinetic model of E. coli:

• The iML1515 model from BiGG Models is a genome-scale metabolic model of Escherichia coli str. K-12 MG1655. It includes 1516 genes, 2712 reactions, and 1877 metabolites, representing a comprehensive reconstruction of the organism's metabolism. This model serves as a robust tool for simulating metabolic fluxes, predicting growth rates under various conditions, and exploring genetic modifications for biotechnological and research applications.
• The E_coli_Millard2016.xml model captures the main central carbon pathways and accounts for 68 reactions and 77 metabolites which are located in 3 compartments: environment, periplasm and cytoplasm. The model represents glucose-limited conditions and is expressed as a set of ordinary differential equations.

Both models were modified adding the reactions necessary for citramalate synthesis (see paper "Enhancing Genome-Scale Metabolic Models with Kinetic Data: Resolving Growth and Citramalate Production Trade-offs in E. coli"). The included files and scripts were used to analyze the updated apply the information of the kinetic model to the genome-scale metabolic model in both situations (original and citramalate-producing scenarios). We evaluated flux metrics such as dormant reactions, flux variability, and other key parameters.

Only the steps of the procedure for the non-producing citramalate model (original) are displayed but the same process should be applied for the "citramax" directory.

1. metrics_fun.py: Includes utility functions to:

   • Count dormant reactions.
   • Assess uncertainty (flux flexibility) after applying kinetic constraints.
   • Generate cumulative distributions.

2. graphs_nocitra.py: Runs simulations and generates visualizations for:

   • Dormant reactions vs. the number of implemented kinetic constraints.
   • Variability in reaction fluxes.
   • Comparisons between original and kinetically constrained models.
   • graphs_with_citra.py is utilized in the citramalate-producing scenario.

   The output shows the FBA solutions after the implementation of the kinetic bounds of each of the mapped reactions. Each solution represents a point in the output graphs.

3. core_simulation.py: The core simulation script. It integrates the iML1515 model and applies kinetic constraints based on the the steady state simulation of the kinetic model. It includes functionality for:

   • Flux Balance Analysis (FBA) optimization.
   • Flux Variability Analysis (FVA).
   • Iterative implementation of kinetic constraints.

4. Metabolic Model: The model used is iML1515.xml. Replaced with iML1515CITRA.xml in the citramalate-producing scenario.

1. Python Requisites:

   • cobra
   • numpy
   • matplotlib
   • pandas
   • xlwt
   • xlrd

   Install them using:

   pip install cobra numpy matplotlib pandas xlwt xlrd

2. Kinetic Flux File: Include an Excel file (kinetic_fluxes.xls) containing kinetic flux values used to apply constraints.

Ensure the metabolic model file (iML1515.xml) and kinetic data file (kinetic_fluxes.xls) are in the same directory as the scripts.

Run the main script to simulate metabolic optimization:

python3 graphs_nocitra.py

This script:

• Calls simulation functions in core_simulation.py.
• Produces visualizations comparing results between the original and kinetically constrained models.

The following graphs are generated:

• Cumulative distribution of flux variability.
• Number of dormant reactions after applying constraints.
• Comparison of growth rates and metabolic variability.

1. Metabolic Model:

   • The iML1515 model is used, representing E. coli's metabolic reactions.
   • A fixed glucose flux is applied as a substrate.

2. Kinetic Constraints:

   • Kinetic fluxes are derived from the kinetic model in E_coli_Millard2016.xml which was simulated until steady state was reached giving the fluxes of kinetic_fluxes.xls
   • Lower and upper bounds are applied to key reactions.

3. Results Analysis:

   • The impact of kinetic constraints is evaluated on dormant reactions and flux variability.

Before running the simulations, it is essential to have installed a linear programming solver. In this work we used CPLEX, but others such as GLPK and Gurobi are supported as well.

To easily install GLPK:

     sudo apt-get install python-glpk
     sudo apt-get install glpk-utils