20.320  Analysis of Biomolecular and Cellular Systems

The Final OH

Hey class,

It's been a pleasure serving as your TAs this semester and because we like you so much, we are going to have one final office hour tomorrow (12/16) from 5-6 PM in usual room (56-614). Come to ask questions, panic (hopefully not), get that final clarification, or just chat.

Hope all of your studying is going well!

Best,

The 320 TAs

P.S. If you have questions now or in the future outside of the class, science or career related, we would love to help out. Feel free to send us an email anytime.

Announced on 15 December 2013  3:30  p.m. by Ryan Lewis Kelly

Final Exam Review Session

Dear Class,

We will be having our final exam review session this Friday at 3 pm in 56-614. Please come with many questions! There may also be a delicious surprise for you!

It has been a pleasure serving as your 20.320 TAs, and we hope you learned and gained a lot from the class. Thanks for being great students!

Hope to see you Friday!

Cheers,
Marianna+Ryan+Bo

Announced on 10 December 2013  11:34  p.m. by Marianna Sofman

Important: Design Project Detailed Guidelines

Hi class,

As the deadline for the project approaches, I thought we would give you a detailed list of what we expect in your final reports. This should make things easier for you and for us as well.

Overall:

To turn in:
Physical to box: Full report with graphs (no code); Digital: Full report in .pdf format, all .m files used to generate figures (we should be able to hit run once in the main function to produce all figures)

You only need to answer the questions given, the first thing we should see is Part 1a, a schematic of your system.

All figures should be clearly labeled with units on axes and legends.

Part 1: Conceptual Design (12 pts):
Please provide a clear schematic of your system design along with an explanation of how it works and why it is biologically feasible

Part 2: Mathematical Design (38 pts):
a-d) Provide clear listings of ODEs and parameters, along with any assumptions that went into them. Also justify why your parameter values were chosen as is.
e) Please generate 2 figures, each with 3 subplots. These should be:

Figure 1: System response to Low Glucose (50 nM) in transient (5 min) and persistent (60 min) cases. Your input should have the following characteristics: Off 5 min, on 5 min, off 10 min, on 60 min, off 10 min. Subplot 1: Glucose input. Subplot 2: all state species. Subplot 3: Insulin and Leptin outputs

Figure 2: Same as figure 1, but for high glucose (10uM)

e-f) Clear answer to questions with explanation

Part 3: Sensitivity Analysis (20 pts)
a) Pick three characteristics of your system, and state their relevance
b) Pick 2 parameters for your system, and for each parameter do the following:
1) Change the parameter by at least 10 fold above and below the final design parameter (3 conditions minimum, eg 10x, 1.1x, 0.1x)
2) Evaluate your 3 characteristics from part A in each of these new parametric regimes.
3) Calculate the relative deviation in response for each change (S, from supplemental info). Note in formula: should be ((Pnew-Porig)/Porig)/((Kmod-Korig)/Korig)

You can either convey this information in a table or in a figure, as seen in the supplemental info.

c) Pick one parameter that speeds up the system (faster time to steady state), plot the high glucose response as in Part 2 and comment on how this speeds up response

d) Comment on whether or not changes in initial conditions affect your system; you may want to include a representative graph to support your argument but not required

Part 4: Robustness (15 pts)
a-c) Use the supplied code and try to "break" your system. Include a representative output (high glucose response) for the noise/chatter/amplitude that is able to disrupt your system. If you are not able to disrupt the system, comment so and include a graph at the highest amplitude tested.

Part 5: Future Work (10 pts)
a-b) Straight forward, answer the questions asked

Misc: Presentation (5 pts): Make it neat and clean!

Also a few comments on output:
1) Having a perfect system is not essential to getting a high grade, we more care that you understand your system and where it works and breaks down.
2) Many have asked or spent a great deal of time getting the relative magnitude of the High/Low responses exact or the time delay exactly 30 min. We care more that the system shows responsiveness to [Glu] and has a time delay in the ball park of 30 min, not that it is exact. If you are close, please go ahead and move on, your system is just fine.
3) Get started on sensitivity early, this will likely require extracting SS values in various parametric regimes.

Best of luck in the remaining few days of the project, we will be at OH today from 3-4 and tomorrow from 7-8. We sometimes will stay late to make sure you all have excellent projects. After the OH at 8pm on Thursday, please do not expect any email responses, so try to get going and work through your problems before then.

Best luck and you're almost there!

-The 320 TAs

Announced on 04 December 2013  10:40  a.m. by Ryan Lewis Kelly

Design project: clarifications on timestep and input magnitude

All,

***** Timestep *****
The current definition for timestep is that each row in sensorInput corresponds to 500 time units, where the unit is the same unit for time as you defined for your rate constants. E.g. if your rates have units for time in seconds, current definition would mean that each row of sensorInput correspond to 500 seconds. The value 500 is defined in tStep on line 109 in sensorODE_solver.m.

If you a conversion factor such as 100 time units = 1 min, this would implicitly mean that all of your rate constants for time would be in terms of centi-minutes and if tStep=500, each row of sensorInput corresponds to 5 min. This is fine.

Optionally: we are allowing you to modify tStep if you'd like to other values. If all your rate constants have units for time in seconds and you change tStep to be 300, this would mean then each row in sensorInput correspond to 5 min.

Note: you still should not modify anything else in sensorODE_solver.m unless you found a bug we did not catch. If so, let us know.

***** Input (glucose conc.) magnitude *****
maxInputAmpl is the multiplier used for your input. If maxInputAmpl = 10, the sensorInput value is multiplied by 10. You can use this later to test the robustness of your system. (Technically sensorInput was meant as boolean matrix with 0s and 1s, and use maxInputAmpl to define your input magnitude, but this is not strictly enforced in code in sensorODE_solver.m so you don't have to do it this way).

Best,
20.320 TAs

Announced on 01 December 2013  6:24  p.m. by Boyang Zhao

Additional OH this week

Hi class,

Instead of recitation this Monday we will be holding an additional office hour at the normal time and place (4-5, 32-124). We will also hold our normal Tuesday/Thursday 7-8 and Wednesday 3-4 OH.

Hope you all are having a great Thanksgiving weekend and best of luck with the project!

-The 320 TAs

Announced on 01 December 2013  11:26  a.m. by Ryan Lewis Kelly