Team:Edinburgh/Notebook/Modelling

From 2010.igem.org

Revision as of 12:37, 25 October 2010

Modelling

14/6/10

First meeting regarding modelling capabilities, decisions of what software to use and other priorities.

17/6/10

Presentation of basic ideas to advisors. Given the green light to proceed using tools and languages discussed.

18/6/10

Commencement of detailed research regarding proposed pathways and their actions.

21/6/10

Setting up the various tools required to use Kappa throughout the project. Collaboration with Donal's Masters project, and the simulation package he's developing for it. Debugging and testing of said package to continue throughout project timeline.

22/6/10

Commencement of analysis and understanding of Ty Thomson's repressilator model, as well as the underlying framework proposed for BioBricks. BioBricks are not characterised to the level of detail necessary for full accuracy, so apparent even at this stage that the initial model will focus on design and prototyping of the system rather than precise interactions. Presentation of framework as a possible future standard of characterisation?

Also talked to biologists regarding the model itself and what assumptions can be safely made, i.e. levels of light linked to expression levels of light-producing protein, in turn indirectly linked to the characteristics (such as transcription rates) of promoters. The ultimate modelling question could thus be: how to calibrate production of protein (and thus light) such that the sensor is able to react to it appropriately? Will we be able to characterise the parts to the correct extent such that we're able to do this accurately?

... we hope so.

24/6/10

Plan of action drawn up for the next two weeks, including the proposed milestones for the modelling side of the project. Aim is to get preliminary model done (and debugged, very important this) by 10/7.

Repressilator model tweaked slightly for our purposes. Preliminary modelling of red and blue light pathways (both based on pre-existing parts, unlike the proposed green pathway) begun.

3/7/10

Base pathways in place. Debugging phase commenced. Bring out the sledgehammers!

6/7/10

Two pages worth of process of elimination later, critical 'no oscillation' bug found to be caused by software issue rather than modelling error. Now resolved. Take that, Murphy!

7/7/10

Balancing of rate parameters in order to elicit desired behaviour. A very delicate process, especially given that mostly working blind without much wet-lab characterisation / quantification data. Rates are mostly guesstimates (and restricted such that they don't crash either the lab computer or the laptop through the presence of Far Too Many Agents).

Issues involve copy number assumption, the action of the OmpF and OmpC promoters, the promoter mechanism of the repressilator, choice between indirect action (i.e. RED light on Cph8) vs. direct action (phosphorylation of OmpR), and so forth. Hopefully I've actually managed to document all the assumptions and decisions that I've made!

12/7/10

Preliminary model complete. All modelling work now put on hold while wiki design and implementation takes centre stage, also to try to allow wet-work to catch up.

16/8/10

Have decided that Kappa is a lot less soul-destroying than CSS. At least the wiki looks fairly good for now.

Work on second (genomic) model commenced, codenamed 'BRIDGE', mainly to refresh my mind on perturbations and to test the capabilities of the simulator. All previous work now renamed 'Moodlighting', will be revisited soon.

20/8/10

BRIDGE model complete, simulator has been thoroughly debugged and will now serve its purpose for more strenuous perturbation analysis. Have also begun write-up of modelling BRIDGEs on wiki, including an introduction to Kappa.

23/8/10

Perturbation analysis of individual light sensing pathways begun. Basically, this involves controlled perturbations on the isolated pathways - what does the pathway do when unresponsive, and how does it respond to a burst of light from the correct spectrum? Ultimately, it is hoped that we would be able to calibrate the reaction of the pathway to actual experimental results, but with the problems that have been experienced with parts from previous years, it's unlikely that we'll be able to get that far within the project lifespan. Bother.

24/8/10

Found one or two minor bugs that somehow slipped the net earlier, mostly involving transcription factors that deactivate when they're not supposed to. Now fixed.

27/8/10

Perturbation analysis of red and blue light pathways complete. Pleased to say that they work as intended, although how accurately they reflect the real biology is still up for debate.

30/8/10

Discussed plans regarding the creation of an intercellular model detailing how the bacteria will hopefully react to one another or to a mechanical light source, as opposed to the intracellular model which we've been concentrating on so far. This may or may not involve a composite model consisting of individual agents (cellular automata or Kappa model) each reacting as dictated by the intercellular model... mindboggling. Decided to start out simple and work our way upwards.

Also discussed plan of action for further perturbation analysis of the intracellular model. This will be conducted and written up within the next month.

1/9/10

Began running preliminary tests on the system as a whole. The results look promising!

6/9/10

Began integration of hypothetical green light producing / sensing pathways into model.

8/9/10

Completed integration of green light producing / sensing pathways into model. Now perturbation analysis of the intracellular model can begin in earnest.

10/9/10

Simulations progress throughout the day. Data is being collected. Mmm... data...

Also have a detailed outline for the intercellular model. There's still a few kinks that need to be ironed out, but with any luck...

30/9/10

The last of the simulations has been completed. Analysis of the data obtained is promising.

The intercellular model, on the other hand, is proving to be very troublesome - the complicated nature of the project is making it difficult to implement the necessary code.

7/10/10

The last of the analysis of the intracellular model is complete and has been written up on the wiki.

Troubles still continue with the intercellular model. The good thing is that we'll be able to concentrate upon it fully now, the bad thing is that there's not much time left. With luck, we should be able to get at least some results from an initial set-up... hopefully...

22/10/10

Good news regarding the intercellular model? Results will be up on the wiki by the date of the freeze...