igemsoftware2020 / hannover Goto Github PK

I want to be more flexible at switching between different bacteria types and the associated constants. as well as simulation constants
Therefore the constants class will need a rework to be able to CRUD constants. This will also come in handy, when building a GUI to manage different simulation runs.

To make the application even easier, I want to merge all of the classes in a python package. The goal is that the model can be installed using pip.

Research cell adhesion forces to accurately model cell cell adhesion and cell- petri dics adhesion.

The animation starts at the end and not at the beginning of the simulation. Tried to revert the input lists in the animate() function in plotting.py but this didn't fix it.

If the time step is set unequal to 1 second, the step size of the time axis in the plots will be wrong.
Fix this by adding the Constants as a parameter of the respective plotting function.
Then call constants.get_simulation_constants(key="time_step") and multiply the step size with it.

I want to take the exported simulation parameters of the bacteria and store them as a json. Currently they are exported in the format:

#********************************************************************************************

#coding:
#index, totalForce_equivalent, living, moving, width, length,
#[pos[0], pos[1], pos[2]], [velocity[0], velocity[1], velocity[2]],
#[angle[0], angle[1]], [velocity_angular[0], velocity_angular[1]];
#********************************************************************************************

frame 0
1; 0.0; True; False; 6.0; 483.9587737429336; [72.42471235743828, 72.21358890260774, 13.016747659390617]; [0.0, 0.0, 8.20055102541609]; [3.0031802567586254, -0.7060105708398637]; [-0.0007403985920757826, 0.032618881148861176];

frame 1
1; 0.0; True; False; 6.0; 487.92840095914994; [72.42471235743828, 72.21358890260774, 26.60554760144548]; [0.0, 0.0, 8.560943963494564]; [3.002444613086532, -0.6311192925981083]; [-0.0004634555134190287, 0.04718150529230595];

Switching to jsons is maybe better for importing the values and could provide a better overview.

Goal is to read in the exported bacteria parameters for each frame and make a proper visualisation. So we wont have to look only at the 3D modelling video but can actually track the parameter values over time.

Add a repulsive potential to model bacteria interaction. Maybe a good approach is to use the Lennard- Jones potential:

https://towardsdatascience.com/the-lennard-jones-potential-35b2bae9446c#:~:text=A%20Soft-Disk%20fluid%20simulation%20based%20on%20the%20Lennard-Jones,pairs%20of%20particles%20exploiting%20repulsive%20and%20attractive%20forces.

We noticed, that even though multiprocessing is enabled, there is room for improvement. The CPU uses all of the available core but the workload is not at 100 %. This means the simulation can be made even faster.
We want to find out the reason for this.
This is probably a big issue, which requires major changes to the code when implemented.

Bacteria don't always split at a fix value. Rather the splitting length is a gaussian distribution. Adapt the splitting length, which is fix at the moment, to a more variable one.

I want to add a quantative variable for EPS concentration. EPS are biopolymers of microbial origin in which biofilm microorganisms are embedded. For our purposes, the EPS may be thougt of a glue. EPS concentrations should have an influence on drag forces by increasing friction.
An approach may be adopted by https://iwaponline.com/wst/article-abstract/43/6/135/9448 , which use the Luedeking and Piret equation to model EPS production kinetics,

Model substrate concentration over time. The Biofilm will grow in a substrate a will take nutrients from it. Research the influance of substrate concentration on biofilm growth.