pwendering / AraTModel

Integration of temperature-dependent constraints into a metabolic model of Arabidopsis thaliana

Metabolic modeling of temperature effects on growth and metabolism of Arabidopsis thaliana

Requirements

The code was tested on both Windows 10 and Ubuntu 20.04.3 LTS.

To execute the code, the following software must be installed

• Matlab (tested with 2020a/b)
• Gurobi solver (tested with version 9.1.1)
• COBRA toolbox for Matlab
• (only for statistics of growth experiment) R programming language (tested with version 4.1.2)

Setup

Clone this repository by running the following code in the command line

git clone https://github.com/pwendering/AraTModel

Change into the AraTModeldirectory and run config to add the functions to the Matlab search path.

Constraint-based analysis of temperature effects

The function simulateTempEffects.m adds temperature-dependent constraints on

• k_cat values,
• total protein content,
• photosynthesis (FvCB model) and CO₂ uptake,

and solves the resulting quadratically-constrained optimization problem(s).

The function takes two required input arguments:

• TGEM, which a COBRA-formated metabolic model with additional fields required for the addition of temperature-dependent constaints
  • the configured model based on the enzyme-constrained AraCore model can be loaded by running

   load(config('tgemFile'))
  

• I, the light intensity that should be used in the simulation

By default, a simulations will be carried out at temperatures between 10 °C and 40 °C.

Optional input arguments can be given as key-value pairs. For instance, a custom temperature range (20 °C to 30 °C) can be used:

simulationResult = simulateTempEffects(TGEM, I, 'tempRange', celsius2kelvin(20:30))

The function returns up to four results:

• simulationResult: growth rates (mu) and net CO₂ assimilation rates (A) and the underlying flux distributions from with and without minimization of total flux
  • For mu and A, the minimum, average, and maximum values from 20 pool solutions are returned. For the publication, the maximum values were used.
• photParams: temperature-adjusted values that were used to parametrize the constraints based on the FvCB model
• fvcbParams: temperature-adjusted values of obtained by the original FvCB model (Farquhar et al. 1980, 10.1007/BF00386231)
• gurobiProblem: pFBA problem with temperature-dependent constaints formatted for the Gurobi solver
  • this is the problem solved for the last temperature if a range is given as input

Configuration of a metabolic model for temperature-dependent analysis

1. Create an enzyme-constrained model using GECKO 2.0.2 (not compatible with later versions). See script create_a_th_ec_model.m for specific changes introduced to the AraCore model. Specific adjustments made to functions of the GECKO toolbox can be obtained upon request.

2. Adjust config file if necessary

• input model file
• UniProt protein information file
• photosynthesis parameters
• protein content data
• IDs of some relevant exchange reactions
• (predicted) optimal temperatures
  • can be generated using the command line application at https://github.com/pwendering/topt-predict
• file containing TPP data (see protein-stability/ath-protein-stability.csv for the format)
  • if downloaded from the Meltome Atlas, adapt and use code/bash/get-protein-stability-data-ath to create the file

3. Add additional fields using the createTGEM function (see also code/matlab/thermo_model/create_etc_model.m for an example)

Reproduction of published results

All scripts that generated the results presented in the publication are located at code/matlab/analysis_scripts, code/matlab/plotting, and code/R.

Reference