Created as part of my M. Sc. project in the Biomedical Engineering Graduate program at the University of Manitoba. Please see the MIT License file for usage and distribution details.
This repository houses all of the code necessary for running my masters project in MATLAB. The code housed here was initially developed in MATLAB R2014a and then further developed in MATLAB R2015a.
Some of the important files for running this model are: main.m and setup.m. Main.m starts the program by first generating all of the parameters and relevant model information by calling setup.m. Next the program is displayed by passing in the parameters structure to finalgui.m, which controls the display of the GUI and the functionality of all of the GUI components.
To edit the system parameters, many of these edits can be carried out directly in the GUI. For more information on editing within the GUI, see the section below on The Components of the GUI. Parameter values, initial values, and data sources are all directly modifiable in the GUI. For changing default values in any aspect of the model, however, they must be edited at the time of creation of the parameters structure, found in the setup.m function.
As detailed above, the model can be run simply by navigating to the location of main.m and then running this in MATLAB. The model will be initialized and the GUI will be promptly displayed. Before any simulation is run, the parameters and initial conditions are set to default values. To change these before simulation, they can be directly edited in the text boxes that accompany the value desired. To run a simulation of mitochondrial bioenergetics function, hit the Plot button in the bottom section of the GUI. This will solve the model using ode45 or ode23t in MATLAB, and then plot the resulting quantities for all four sections of the graph.
In this figure, the GUI layout is shown, including axes and all input areas. With respect to input areas, the top three sections control:
- Initial Conditions
- The user can vary the input to each of the listed initial concentrations
- There is a randomize button for generating a random set of initial conditions
- The keyboard hotkey for the randomize button is R
- There is a default button for resetting the initial conditions to default values
- The keyboard hotkey for the reset initial conditions button is E
- Parameter Values for a Control Condition of the Model
- The user can edit each parameter value for the calibrated model, although this set of options is by default disabled
- To enable these settings, click the radio button labelled Enable Editing
- Parameter Values for an Experimental Condition of the Model
- The user can by default freely edit each parameter value directly, which will update the model values
- There is also a default reset button found in the bottom-left section labeled Optimization
- The keyboard hotkey to trigger the reset parameters button is D
Next, the botttom three sections control:
- Optimization
- This button is for running launchQubist.m, which will run Jason Fiege's Ferrret evolutionary algorithm, as part of the Qubist Global Optimization Toolbox, for the purpose of optimizing the parameters to fit given data.
- The keyboard hotkey for the optimization button is O
- Important: This project does not come packaged with Qubist, since it is licensed software for which permission was acquired for use in this project. A self-produced, although less robust, implementation of a genetic algorithm for optimization is under development for public use. If one wishes to use Qubist or are interested in learning more about this Toolbox, please visit http://www.nqube.com/
- The load params button allows the user to search for a -BestResults.mat file, which is automatically
generated by Ferret using the post-processing script analyzeResults.m. These files are housed within the
/DeterministicModel/Solutions/ directory.
- The keyboard hotkey for the load parameters button is L
- There is a default button found here, as detailed above, for resetting the parameter values in the experimental
condition
- The keyboard hotkey for the reset parameters button is D
- This button is for running launchQubist.m, which will run Jason Fiege's Ferrret evolutionary algorithm, as part of the Qubist Global Optimization Toolbox, for the purpose of optimizing the parameters to fit given data.
- Data Source Control
- Here there are three buttons for indicating where data will be imported from, allowing the user to decide whether the model will try to emulate Seahorse XF data, Oroboros Oxygraph data, or from a matlab .mat file
- Plotting
- The user can click the large Plot button to instruct the program to solve the model, and then display the
resulting behavior of the system in the axes found to the right of the control panels.
- The keyboard hotkey for the plot button is P
- The user can click the large Plot button to instruct the program to solve the model, and then display the
resulting behavior of the system in the axes found to the right of the control panels.
The axes illustrate the behavior of the system once Plot is clicked. As shown below, once it is selected, the system then graphs and labels each axis to display what the model is simulating with regard to mitochondrial function.
Several additional functions were introduced into the GUI of the program, each of which will be discussed in detail in this section.
Each graph, once plotted, can then be right-clicked for additional options. These options allow for the user to either save a blown-up image of the plot to a PNG file, or for opening the image in a new figure window for inspection and editing purposes. Both options facilitate creating large images of the graphs for presentation, publication, or documentation.
Opening the graph in another figure window enables editing and formatting within the figure MATLAB editing environment. Labeling can be controlled using the figure menus built-in to MATLAB, and the figure can be exported into any format after editing is finalized. Zooming in and out of the graph can also be done in this figure window.
In the File menu, there is an option for saving a snapshot of the GUI. The purpose of this is to save the current display of the GUI either for presentation or publication, where the graphs of each element of the system are shown along with the set of parameters and initial conditions that were used to acquire this behavior. Default location for saved images are in the StateImages directory.
The keyboard shortcut for this function is ctrl+d on windows and linux systems (with windows keyboard layout, not EMACS) and cmd+d on MAC OS X systems.
In addition to being able to save a snapshot of the GUI as an image to be viewed or presented later, the state of the model can also be saved. The save and load session commands are found in the File menu, with keyboard shortcuts ctrl+s for saving a session and ctrl+o for opening (or cmd+s and Cmd+o on OS X).
Saving a session allows the user to save the current values of all variables and all graphics objects in the guidata and handles structures, which will default to save as a .mat. Once this .mat is saved, it can later be reloaded using the load command giving the user the chance to re-open the state of the chosen file. Saved sessions are generally stored in the Savestates directory. Make sure to only load session .mat files.
In the Help menu, there is a Version button and an Info... button. The Version button allows the user to confer with the Git system to check the current tagged version of the program. This will inform the user whether or not there are any available updates to the model or the GUI code so that the most up-to-date version of the simulation model can be accessible to the user without complication. The Info... button will pull up this README.md document in the appropriate default program.