Title: Current Methods of Characterization
1Rapid Characterization of Cellular Pathways Using
Time-Varying Signals Ty M. Thomson and
Drew Endy Division of Biological
Engineering, MIT Contacttmt_at_mit.edu
Input Microfluidics
Biological System Yeast Pheromone Signal
Transduction Pathway
A microfluidic device was designed, and
subsequently manufactured out of PDMS, to allow
for time varying stimuli to be applied to cells
anchored in a channel. On-chip valves allow for
rapid control of the flow rates of two fluids
through a single channel. By varying the flow
rates, the position of the boundary layer between
the fluids is altered, exposing cells anchored in
the channel to varying extracellular environments.
- The pheromone response pathway is an obvious
model system to use. - It is a well-studied prototype for regulatory
networks that govern response to external stimuli
in higher eukaryotes. - It contains many common elements of signaling
pathways (MAPK cascade, G protein, etc.) - A number of stimuli and reporters are now
available for this pathway, including specific
inhibitors for several kinases.
0s
Current Methods of Characterization
0.1s
0.2s
Time Varying Stimulation - Parameter Sensitivity
Response to 1?M pheromone simulated data, and
model results for Gpa1/Ste4-Ste18 dissociation
rate 10x too high.
In general, the output of a given experiment or
simulation will depend critically on certain
parameters, and depend weakly on others. Fitting
parameters in our model to experimental data
gives us strong data for some parameters (on
which the results critically depend) and weak
data for others. We are using a computational
model of the pathway to identify stimulus
time-courses that have the greatest potential to
produce highly informative experimental results.
0.3s
Same as above plot, but data is linearly scaled
to match simulation (since data in arbitrary
units).
0.4s
Response to two 1?M pheromone pulses (at 0-10s
and 190-200s) simulated data (scaled), and model
results for Gpa1/Ste4-Ste18 dissociation rate 10x
too high.
For several stimulus time-courses, quantify mean
square error
Time-Varying Input/Output
Biological System
Output - Reporters
Input ?-fluidics
Output Signal (Experimental)
Acknowledgements
Conclusions
- Kirsten Benjamin
- Alejandro Colman-Lerner
- Larry Lok
- Jeremy Thorner
- Todd Thorsen
- Endy Lab
- Molecular Sciences Institute
- Alejandra Torres
- Numerica Technology John Tolsma
- Time Varying stimulus time courses show a
definite improvement in parameter sensitivity
over a step input, which should improve our
estimation abilities for some parameters. - Time-varying stimulation of a pathway will likely
not increase sensitivity enough for some other
parameters to allow for accurate estimation - Controllable microfluidics is a practical method
for controlling the fluid environment of
immobilized yeast cells on sub-second timescales. - My technology platform will improve and scale
along with advances in fluidics, reporter
technology, and hypothesis testing and non-linear
parameter estimation as they pertain to cellular
systems.
Input Signal
Refine Model
References
Gpa1/Ste4-Ste18 dissociation rate
Pheromone/Ste2 dissociation rate
- Microfluidic Soft Lithography Foundry -
http//nanofab.caltech.edu/foundry - Ficarro, S. et al. (2002) Nat. Biotech. 20,
301-305
Output Signal (Simulated)
Computational Model
Future Directions
Support
Signal Design
- Use microfluidics to stimulate the system with an
information-rich, time-dependent signal in order
to observe more varied pathway behavior. - Measure the states of various reporters over
time. - Evaluate hypotheses by comparing computational
results with experimental results. - Test inferences and improve understanding by
selecting a new input signal and repeating the
process.
- Further characterize cell behavior in the
microfluidic chip environment. - Begin to perform pheromone response experiments
with chip.
- NHGRI Center of Excellence in Genomic Science
- CSBi Cell Decision Process Center
- MIT Presidential Fellowship