Minnesota/23 June 2009

From 2009.igem.org

(Difference between revisions)
 
Line 5: Line 5:
|'''[[Minnesota/22 June 2009|Go to Previous Day (June 22)]]'''|| width=158|'''[[Minnesota/24 June 2009|Go to Next Day (June 24)]]'''
|'''[[Minnesota/22 June 2009|Go to Previous Day (June 22)]]'''|| width=158|'''[[Minnesota/24 June 2009|Go to Next Day (June 24)]]'''
|}
|}
 +
'''Patrick'''<br>
 +
A workaround was discovered for the non-linear parameter variation issue mentioned before. Basically, you create a linear variation ([0 1 2 3 4 5] for example). Then you can use ncdump and ncgen, two programs on the supercomputer and available to our MATLAB package from MSI, to change these numbers to the appropriate (non-linear) values. I quickly created two MATLAB m-files to do this automatically for a given .nc-file automatically, because altering the nc-file in the supercomputer was too time-consuming.<br>
 +
 +
I've now begun cleaning up the reaction network so I can be sure everything is running as I want it to.
 +
 +
'''Ben'''<br>
After submitting the model from yesterday, we got the results from the supercomputer.  
After submitting the model from yesterday, we got the results from the supercomputer.  
We compared the gfp that was produced by the model to the experimental results that were found by other members of our group.
We compared the gfp that was produced by the model to the experimental results that were found by other members of our group.

Latest revision as of 21:19, 29 July 2009

Back to Notebook Home
Go to Previous Day (June 22)Go to Next Day (June 24)

Patrick
A workaround was discovered for the non-linear parameter variation issue mentioned before. Basically, you create a linear variation ([0 1 2 3 4 5] for example). Then you can use ncdump and ncgen, two programs on the supercomputer and available to our MATLAB package from MSI, to change these numbers to the appropriate (non-linear) values. I quickly created two MATLAB m-files to do this automatically for a given .nc-file automatically, because altering the nc-file in the supercomputer was too time-consuming.

I've now begun cleaning up the reaction network so I can be sure everything is running as I want it to.

Ben
After submitting the model from yesterday, we got the results from the supercomputer. We compared the gfp that was produced by the model to the experimental results that were found by other members of our group. Here are the graphs:
Tnnatcvaries4.jpgTnnexperiment.jpg

From these graphs, it is possible to see that the model is not really very correct. There is no GFP production at an aTC concentration of 0, when the experiments show a lot of production. This phenomenon is called leakiness and it is something that will need to be dealt with in future models.
Also, the model seems to peak at 3 hours or so, and then slowly decreases. This is not what we want to see. Instead the model should peak at 6 hours and then level off to a steady state.
The next model will hopefully improve on this effort and fix some of the problems.