Team:Wageningen UR/Notebook/Journal/Kinetics

• Project

• Background

• Wet-lab

• Results
• BioBricks

• Model

• Cpx Kinetics
• Quorum Sensing
• Cpx Integration
• QS Integration

• Human Practices

• Integrated HP
• Safety & Ethics
• Collaborations
• Future Perspectives
• Engagement
• Meetup

• Application

• Demonstrate
• Device
• Roll-out Plan

• Notebook

• Protocols
• InterLab
• Lab Journal
• Achievements

• Team

• Members
• Attributions
• Acknowledgements

• 
• Project
• Background
• Wet-lab
• Model
• Human Practices
• Application
• Notebook
• Team

• Wet-lab
• Model
• Application
• Human Practices
• Notebook
• Team
• Parts
• Results
• Cpx Kinetics
• Quorum Sensing
• Cpx integration
• QS integration
• Integrated HP
• Safety & Ethics
• Collaborations
• Future Perspectives
• Engagement
• Meetup
• Demonstrate
• Device
• Rollout Plan
• Protocols
• InterLab
• Journal
• Achievements
• Members
• Attributions

• Home
• Notebook
• Journals
• System Kinetics

Modeling System Kinetics

By: Sabine van Oossanen

Here you can find Sabine's work on the model describing the Cpx system. Curious about her results? Read more about it here.

May

Week 1 (1st of May - 7th of May)
Learning Matlab and Modelling basics through a short Maths lecture and multiple exercises.

Week 2 (8th of May - 14th of May)
Finished thesis proposal. Discussed with receptor and BiFC students on system layout, finally decided on three different ways in which the system could work. Aim of the thesis is to compare the systems kinetics, sensitivities and their maximum output.

Week 3 (15th of May - 21st of May)
More system design and learning Matlab (and getting used to my Mac).

Week 4 (22nd of May - 28th of May)
Started on the second model layout (Setup 2), which involves GFPc bound to CpxA and GFPn to CpxR. Set up reaction equations (ODE’s).

June

Week 5 (29th of May - 4th of June
Improved the reaction equations and started writing in Matlab.

Week 6 (5th of June - 11th of June)
First runs of model. The model gives the concentration of each compound over time, given a set op reaction speeds (parameters). As the parameters of most reactions of the system are not known, parameter optimisation can be used. Here, random parameters were generated using Latin Hypercube sampling. The objective used for selecting the best parameter sets was a maximal final GFP concentration that could be formed with those parameters. Learned how to introduce pulses to the system.

Week 7 (12th of June - 18th of June)
Improved parameter generation script. Generated 100,000 parameters sets, still found too much inconsistency when analyzing the best 10%. Retried with 500,000 parameter sets and learned about lognormal distributions. Included parameterset distributed by log-uniform distribution.

Week 8 (19th of June - 25th of June)
Learned about servers, terminal, screens and how to generate parameters outside the Mac. This is to run processes continuously and parallel to each other.

July

Week 9 (26th of June - 2nd of July
Developed second objective score script, aimed at saving the time it takes to reach half of the final GFP concentration. This was tested using setup 2 data. The log-uniform distribution and uniform distributed sets were compared based on speed and GFP scores. These were not the same.

Week 10 (3rd of July - 9th of July)
This week was spent on figuring out and selecting the pareto sets of the setup 2 parameter set scores.

Week 11 (10th of July - 16th of July
As the speed score itself is not very characteristic for the efficiency of the parameter set, a new score takes both time and increased GFP concentration into account. Some conclusions were made on the characteristics of good Setup 2 parameter sets.

Week 12 (17th of July - 24th of July)
Start setting up Setup 1, in which dimerization of phosphorylated CpxR leads to GFP maturation. Log-uniform distributed sets were generated.

Week 13 (25th of July - 30th of July)
Analysis of setup 2 was finished. Conclusions were made on setup 1. Together with Bart and the supervisors, possible future experiments were discussed to combine wet-lab with the model.

August

Week 14 (1st of August - 7th of August)
Start of setup 3. New sets were generated. Setup 2 was adapted to conform with setup 3, as this does take into account background GFP formation. New scoring function taking background GFP was developed.

Week 15 (8th of August - 14th of August)
Holiday - Out of office!

Week 16 (15th of August - 21st of August)
Start of sensitivity analysis. This week I set up the draft script in which the model is simulated under different concentrations of antigen. I thought of several ways of visualizing the influence of the antigen concentration on the objective scores.

Week 17 (22nd of August - 28th of August)
Go/No-Go presentation! In addition I extended the sensitivity analysis to include varying concentrations of CpxR and CpxA, as these are the key players in producing the YFP signal.

September

Week 18 (29th of August - 4th of September)
Looking critically at the setups 2 and 3, I saw these could be simplified. Thus, I simplified the setup 2 and compared this to the behavior of setup 2 and 3. Additionally, I altered the speed objective score. In the new scoring function, the model is simulated in absence of antigen. At a certain moment, antigen levels are set at a concentration of 1. The GFP levels before and after addition of antigen are taken into account. The speed score is calculated over the time it takes to reach half of the GFP increase after addition of antigen.

Week 19 (5th of September - 11th of September)
The new scoring system was implemented for the setups 2 and 3. Future directions of the Cpx project were discussed with Bart.

Week 20 (12th of September - 18th of September)
The new scoring system was implemented for setup 1 and its effects on setup 2 and 3 were assessed.

Week 21 (19th of September - 25th of September
When applied in the field, Mantis will still have to perform under diverse circumstances. To simulate this effect on the setup performance and see which setup is most robust, I want to compare ideal sets to bad sets. This week I came up with several ways to model this effect.

October

Week 22 (26th of September - 1st of October)
I finished the setup for the robustness analysis and performed it for all three systems. Bart generated data in the lab, I setup a script to compare this data to my setup 1 model and found the settings of the in vitro system.

Week 23 (2nd of October - 8th of October)
The in vitro data was fitted to the setup 1 model, and several parameters were found which gave a similar response. Together with Bart and the supervisors, we discussed what future tests could be done in the wet and dry lab. Additionally, Bart tested a CpxR knockout strain and found that the effect of induction was different for both systems. To find out if this effect is due to the workings of setup 1, I set up a model of the native system which includes native CpxR.

Week 24 (9th of October - 15th of October)
The model of setup 1 in the native strain was compared to the simplified knockout model of setup 1, at first glance this did not explain the effect seen in the lab. Furthermore, to find out which parameters should be changed in the in vitro version of the setup 1 strain, I wrote a script to assess the effect of parameter variation on the in vitro system, given the best fitting parameter set. Several important parameters were found (such as k3 and k4).

Week 25 (16th of October - 22nd of October)
More analysis was performed on the model including native CpxR. The behavior of the native model and knockout model was confirmed to be similar. Possibly the effect seen in the wet lab experiments was due to external factors that was not taken into account in the simplified model.

Week 26 (23rd of October - 29th of October)
Wiki writing on the Human Practices pages, my own results page and many more.