Literature research on pDusk and pDawn system as well as mazEF toxin antitoxin system.
Week 4
Getting familiar with Matlab and setting backbone model structure
July
Week 5-6
First dynamic model structure of pDusk and pDawn was built. Light unit conversion was is given here:
h = 6.63*10-34 ([Js] Planck's constant)
c = 2.988*108 ([m/s] speed of light)
lambdanm = 470 (% [nm] wavelength)
NA = 6.022*1023 ([mol-1] Avogadro's constant)
Ninterm = Ninitial*10-2 ([µW*cm^-2] → [W*m-2])
Ep = h*c/lambdanm (distinct energy quanta of a photon)
Np = Ninterm*lambdanm*106/(1.988*10-16) (number of photons; lambda in [nm])
Eq = Np/NA (photonflux N = Eq in [µmol*m-2*s-1])
N = Eq*60*60 ([µmol*m-2*s-1] → [µmol*m^-2*h^-1])
Week 7-8
Vacation
August
Week 9-11
Improved the equations for pDusk and pDawn. And added the equations for mazE and mazF.
Week 12
Got first useful model results for pDusk and pDawn individually. At this point, there was no weighted means approach.
September
Week 13-14
Realised, that we should combine both the pDusk and pDawn scoring function with a weighted means approach. Started to rewrite the whole script.
Week 15
Vacation
Week 16
Noticed problem with the complex formation. Read up on the complex formation of mazEF. There are actually several stages of the complex. After some iterations, we concluded to model the complex as done in the current equations.
October
Week 17-18
Worked on updating last adjustments. Rewrote all Matlab files to a readable format - noticed quite some small bugs on the way. Ran the analysis and the evaluation. Got some preliminary results from the lab! Implemented that in my analysis last minute.
testing first scripts for a single cell, simulate the model response vs time
Week 3
first results on single cell scripts tested with 10 cells and 10000 parameter sets
expanding model for multi cell population with 10 cells and 10000 parameter sets
testing multi cell, simulate the model response vs time
Week 4
Look at correlations of different parameters in the model. Simulate the model for the GFP response with different parameter sets:, low, medium and high set.
implement positive and negative feedback of LuxR in the system. Test whether the mRNA transcription of LuxR feedback protein has an influence on the system.
June
Week 5
looked at the correlations of the different feedback in the system. To see which parameter has the biggest influence on the system and which parameters could be changed to create subpopulations. Fixed bugs in the system with help of Ronald.
Week 6
With changing the parameters with the biggest correlation different populations can be generated. The LuxR production rate is changed to do so. However, this does not correlate to reality yet. Subpopulations are simulated, but further research is needed to create a system that can be tested in the lab.
July
Week 1
went to iGEM meetup in Paris
Week 3
Expand the old quorum sensing model to 100.000 parameter sets. This needed to run onat the server because it took too long on a single computerthe computer could not handle it. Because the server did not save the work properly it needed to be done a few times. So it took a full week to get results.
Week 4
<p> make the subpopulation system, write the equations and try to get the best parameter sets by searching for a statistical method that does so. For the quorum sensing system.
look at the quorum sensing system with and without feedback.
August
Week 1
Test changes in highest parameter from the quorum sensing system to see if there is a change in GFP production. This is both done for the system with and without negative feedback.
Week 2
Set up the subpopulation system and run it with a few parameter sets. When it worked I changed the concentrations in Arabinose and Glucose in the system. Had some bugs with putting different arabinose and glucose concentrations in the system. Remco helped me solve this.
Week 3
To set up a normal distribution for the quorum sensing part different methods are tried. fixing bugs when making the normal distributions. Plot the amount of GFP vs time. The subpopulation system ran on the server for a couple of days with 100.000 parameter sets. There is searched for the parameter set with the highest RFP, and also for the set with the lowest RFP.
Week 4
Parameter analysis with the best sets, got from the confidence interval. Put the models, quorum and subpopulations together. This did not work out the first time, but the second it did.
September
Week 1
plot normal distributions of the separate systems.
Week 2
look for differences in production of GFP by tweaking the parameters.
Chose the best sets to work with for the combined system. Found four sets that can give the responses we want.
Week 3
Calculate growth rate for the e. coli used by Thomas.
Week 4
Tried the Event function in Matlab to generate oscillations in the system. This failed due to the fact that it could be only used for single cells and not for multi cell populations. We did not know that this was a limitation of the function.
Try to look at the effect of the total volume for a single cell and multicell, but the multicell was not possible.
October
Week 1
I got a code for generating oscillations from my supervisor, but it needed to be debugged. So I spended the week on doing this myself and together with my supervisor.
Make a heatmap for the subpopulation system about the ratio of lambda-cl and 434 compared to the RFP production.
Week 2
look at ratios lambda-cl and 434 to compare with the data Thomas got from the lab. He used a library for the promoter from the subpopulation system. Tried to simulate the combined model with total RFP.
