Sortostat/Notebook

(Difference between revisions)
 Revision as of 10:00, 17 July 2007 (view source)m (→To Do List)← Previous diff Current revision (12:20, 23 December 2007) (view source) (→Analytical model) (14 intermediate revisions not shown.) Line 1: Line 1: - ==Experiments== - *[[Sortostat/Experiments/Two reactor chip]] - - ==Modeling== - *[[Sortostat/Model]] - ==To Do List== ==To Do List== *Figure out the MATLAB memory leak *Figure out the MATLAB memory leak Line 10: Line 4: **Does the leak influence the timing of the events, by making the image processing take too long?  (e.g. why does the timing graph get noisy after awhile?) **Does the leak influence the timing of the events, by making the image processing take too long?  (e.g. why does the timing graph get noisy after awhile?) **Is it the MATLAB thread that blows up or the LabView? it's MATLAB. **Is it the MATLAB thread that blows up or the LabView? it's MATLAB. + **[http://www.mathworks.com/support/tech-notes/1100/1106.html MATLAB help page here might be useful.] *output the direction of sorting to the output file *output the direction of sorting to the output file + *Clean up the Image Processing MATLAB scripts + + ==Done== + *Why do I see any variability in the timing? Timing between cleaning events is rock solid, the timing between sorts is influenced by the time it takes to process the image. + *Change the code so that when flipping based on the CFP it is the moving average that is used, rather than a single image. *Make the list of experiments to evaluate the device performance limits. *Make the list of experiments to evaluate the device performance limits. - *Work on the model and compare to experimental results. + **Sorting at various dilution rates, followed by a sort to extinction. + + ==Experiments== + *[[Sortostat/Experiments/Two reactor chip]] + *[[Sortostat/Troubleshooting]] + ===Future Experiments=== + '''replace bulb before these experiments''' + *check on pH of the media if we are adding AHL, it needs to be <8 for AHL to be stable. + *Zig-zag run where we vary from 0.4 to 0.6 so that we can get more sampling of the selective pressure. + *Maintenance of a slower growing variant vs. a faster growing "mutant". + **Could demonstrate this with a slower growing CFP vs YFP or via something like F2620. + *Zig-zag experiment with different dilution rate (e.g. different total number of cells). + *Get the exact size of sorting chamber by cell counting + *F2620 is in MG1655 whose flagella might cause a problem + **Need to co-transform with RFP expression plasmid anyway. + *We already have the zig-zag automated, but should include a "pulse" as well (e.g. /-\_/-\) where it has regions with no selection. + + ==Modeling== + *[[Sortostat/Model]] *How much faster can we drive it given the timing? *How much faster can we drive it given the timing? *Are there more or less cells in a sort vs a screen? *Are there more or less cells in a sort vs a screen? **No, looked at this before last lab meeting, but should write it up. **No, looked at this before last lab meeting, but should write it up. *Do fluctuations in the total cell count correlate with anything? *Do fluctuations in the total cell count correlate with anything? - *Move all the pictures from the last run to the same HD (back up the ones from the long zigzag run). + *2/18/06 lab notebook page has some information about the appropriate sorting volume to use in the model, move this to the wiki - **Waiting for the new drive to come in. + - *Clean up the MATLAB scripts (particularly the image processing, have lots of old versions lying around). + ===Output=== - *Why do I see any variability in the timing? + *Sortostat mFiles doc - list of all the mfiles that matter (except the image processing).  These are all commented and cleaned up at this point. - *Update MATLAB since the license is running out. + *Sortostat Modeling Probability Distributions doc - description of the modeling of the probability distributions for the number of cells in the sorting chamber and how many are CFP or YFP. + *Sortostat Modeling Stochastic Sim doc - description of the stochastic simulation of Sortostat performance. + + ===Analytical model=== + *put the cleaning period / dilution rate formula derivations on the wiki. + *write up the chemostat equations as they relate to the SS cell concentration (e.g. limiting substrate?  Should we reduce the concentration of glycerol?) + *These come from the SortostatTiming.m matlab model: + **12.5 min between cleaning (doubling time = 0.55) + **13.6 min between cleaning (doubling time = 0.60) + **14.8 min between cleaning (doubling time = 0.65) + **15.9 min between cleaning (doubling time = 0.7) + **17min between cleaning (doubling time = 0.75) + **18.2min between cleaning (doubling time = 0.8) + **19.3min between cleaning (doubling time = 0.85hr) + **20.5min between cleaning (doubling time = 0.9hr) + **21.6min between cleaning (doubling time = 0.95hr) + **22.7min between cleaning (doubling time = 1hr) + **23.9min between cleaning (doubling time = 1.5hr) + + ==LabView== + *why do we need to initialize the load paraemters array? + + ==Methods== + *[[Sortostat/Methods/Receiving chips]] - ==Done== - *Change the code so that when flipping based on the CFP it is the moving average that is used, rather than a single image. ==References== ==References== *[[Sortostat/Foundry report card]] - post notes on each chip fabricated so can document common failure points in fabrication to better avoid them. *[[Sortostat/Foundry report card]] - post notes on each chip fabricated so can document common failure points in fabrication to better avoid them.

To Do List

• Figure out the MATLAB memory leak
• I suspect this is because I'm not closing the image files properly in the MATLAB script that runs, or it could be something more annoying.
• Does the leak influence the timing of the events, by making the image processing take too long? (e.g. why does the timing graph get noisy after awhile?)
• Is it the MATLAB thread that blows up or the LabView? it's MATLAB.
• MATLAB help page here might be useful.
• output the direction of sorting to the output file
• Clean up the Image Processing MATLAB scripts

Done

• Why do I see any variability in the timing? Timing between cleaning events is rock solid, the timing between sorts is influenced by the time it takes to process the image.
• Change the code so that when flipping based on the CFP it is the moving average that is used, rather than a single image.
• Make the list of experiments to evaluate the device performance limits.
• Sorting at various dilution rates, followed by a sort to extinction.

Experiments

Future Experiments

replace bulb before these experiments

• check on pH of the media if we are adding AHL, it needs to be <8 for AHL to be stable.
• Zig-zag run where we vary from 0.4 to 0.6 so that we can get more sampling of the selective pressure.
• Maintenance of a slower growing variant vs. a faster growing "mutant".
• Could demonstrate this with a slower growing CFP vs YFP or via something like F2620.
• Zig-zag experiment with different dilution rate (e.g. different total number of cells).
• Get the exact size of sorting chamber by cell counting
• F2620 is in MG1655 whose flagella might cause a problem
• Need to co-transform with RFP expression plasmid anyway.
• We already have the zig-zag automated, but should include a "pulse" as well (e.g. /-\_/-\) where it has regions with no selection.

Modeling

• Sortostat/Model
• How much faster can we drive it given the timing?
• Are there more or less cells in a sort vs a screen?
• No, looked at this before last lab meeting, but should write it up.
• Do fluctuations in the total cell count correlate with anything?
• 2/18/06 lab notebook page has some information about the appropriate sorting volume to use in the model, move this to the wiki

Output

• Sortostat mFiles doc - list of all the mfiles that matter (except the image processing). These are all commented and cleaned up at this point.
• Sortostat Modeling Probability Distributions doc - description of the modeling of the probability distributions for the number of cells in the sorting chamber and how many are CFP or YFP.
• Sortostat Modeling Stochastic Sim doc - description of the stochastic simulation of Sortostat performance.

Analytical model

• put the cleaning period / dilution rate formula derivations on the wiki.
• write up the chemostat equations as they relate to the SS cell concentration (e.g. limiting substrate? Should we reduce the concentration of glycerol?)
• These come from the SortostatTiming.m matlab model:
• 12.5 min between cleaning (doubling time = 0.55)
• 13.6 min between cleaning (doubling time = 0.60)
• 14.8 min between cleaning (doubling time = 0.65)
• 15.9 min between cleaning (doubling time = 0.7)
• 17min between cleaning (doubling time = 0.75)
• 18.2min between cleaning (doubling time = 0.8)
• 19.3min between cleaning (doubling time = 0.85hr)
• 20.5min between cleaning (doubling time = 0.9hr)
• 21.6min between cleaning (doubling time = 0.95hr)
• 22.7min between cleaning (doubling time = 1hr)
• 23.9min between cleaning (doubling time = 1.5hr)

LabView

• why do we need to initialize the load paraemters array?