BIOLOGY FOR MATHEMATICIANS

Dr. Garrett Odell
Director, Center for Cell Dynamics
Professor of Biology
University of Washington

Monday May 5, 2008
8:00 - 9:00 PM
A Free Public Talk

How, in Silico, the Sea Urchin Embryo Gets its Furrow in the Right Place

Abstract: This talk uses computer-animated movies to demonstrate how an individual agent-based simulation model, involving hundreds of thousands of differential equations solved on computers, constitutes a powerful tool for comprehending how complex cell-level cytoskeletal behavior emerges from molecular-level details of how cytoskeletal proteins interact. To illustrate, Dr. Odell addresses the long-standing cell biological puzzle of how microtubules position the contractile ring that cleaves a cell in two. The first half of the talk presents Dr. Victoria Foe’s recent data on the role microtubules play in furrow positioning. Dr. Foe studied the changing organizations of myosin II and microtubules in fixed and stained urchin blastomeres. Treating blastomeres for five minutes with a high dose of nocodazole unveils a unique population of stable microtubules that forms during anaphase into an array that includes both inter-zonal and astral microtubules, and disappears as abscission completes. Most of these stable microtubules mysteriously point toward the future contractile ring locus. Coincidence between localized phosphorylation (=activation) of cortical myosin and arrival at the cortex of stable microtubule plus-ends implicates these stable microtubules in furrow specification. The second half of the talk uses Dr. Odell's computer simulation model to investigate how MKLP1 (a kinesin motor) moving along stable (but not along dynamically unstable) microtubules can deliver a Rho-activation signal all the way to the cell cortex. It further shows that the presence of myriad dynamic microtubules serves to sharpen the zone of myosin activation, and shows that in order for MKLP1 moving along stable microtubules to accumulate at the future contractile ring site in the face diffusive dissipation, the MKLP1 motors must not walk off microtubule tips (as micrographs confirm they don’t). The embryonic sea urchin egg Dr. Foe discovered the anisotropic array of stabilized microtubules in are huge. Many rounds of division reduces cells to less than one-thousandth the volume of the zygote. Can the mechanism we propose work robustly across such a wide range of cell sizes? The computer model gives the answer.

Tuesday, May 6, 2008
4:00 - 5:00 PM

For Making Genetic Networks Operate Robustly, Unintelligent Non-Design Suffices

Preceding the lecture, tea will be served.
A reception will follow in MSRI's Atrium.

Abstract: Five years ago George von Dassow, Ed Munro, Eli Meir, and Garrett Odell made realistic mathematical models of two ancient and famous genetic networks that act early in diverse embryos to establish spatial gene expression patterns prefiguring the body plan. The models revealed these networks to be astonishingly robust in that they continue to make the correct pattern in the face of thousand-fold variations in the strengths and functional forms of interactions among participating genes. Odell has replaced his initial astonishment to a belief that such robustness is crucial to make networks functionally heritable in polymorphic populations. His computer program haphazardly generates biologically sensible genetic networks which are randomly connected with about the same number of parts as the segment polarity and neurogenic networks. Bjorn Millard, Ed Munro, and Odell wrote computer algorithms that discover and catalog the stable expression patterns any network can make, and, from these, distills those patterns the network can make robustly, with respect to variations of its parameters. We found that 19 out of 20 random networks that our program created could make at least one or many complex stable spatial expression patterns with the same high robustness that the real, evolved networks exhibit.
Their algorithms show that it is possible to replace continuous-time, continuous concentration, differential equation models with quantized far-apart concentrations. Unfortunately, for any given network, there are many different ways to do this. Our in silico result that thoughtless, haphazard, non-design produces networks whose robustness seems inspired begs questioning the capabilities of unintelligent non-design.

Wednesday May 7, 2008
5:00 - 6:00 PM

Why are mindless, individual agent-based computer simulation models less likely to deceive than elegant and thoughtful partial differential equation models of traditional continuum mechanics?

Preceding the lecture, tea will be served.

Abstract: Garrett Odell’s PhD training in theoretical mechanics and mathematics taught him to make and solve classic continuum models (involving mass and momentum conservation laws rendered as partial differential equations) to account for the flow of deformable materials, and he used this approach to try to understand how individual biological cells, or assemblies of cells, moved. In recent times he has abandoned that classic approach and switched to making agent-based models in which each of tens of thousands of individual parts, each governed by a small system of ordinary differential equations, interact with each other. As opposed to the classic continuum model approach in which analytic solutions, or approximations thereto, are possible, these new agent-based models require numerical computer solution of their systems hundreds of thousands of coupled differential equations. At first glance this seems a step backward, away from an elegant, conceptually simple approach in which just the process of formulating the mathematical model distills out a conceptual understanding of what causes the phenomenon under study toward a kind of needlessly complicated and computationally expensive modeling involving more arithmetic than thought. The talk is about why the step described accurately in the line above is a strong step forward rather than backward. He will contrast instances of each kind of model and try to convince you that succeeding in explaining some complex phenomenon (in cell biology for example) using the former continuum modeling approach usually demonstrates only how clever you are while explaining it by the kind of agent-based modeling he has switched to demonstrate that the simple interaction rules by which the myriad parts interact actually can (or cannot) account for what complex phenomena emerge from them. The moral of this philosophy of science talk is that, if you think others should be more interested in what animates cells than in how clever you are, than switch to agent-based models.

LOCATION FOR ALL 3 TALKS
The Mathematical Sciences Research Institute
Simons Auditorium
17 Gauss Way
Berkeley, CA 94720-5070

Parking: Please note that parking in the lot between MSRI and the Space Sciences Lab (SSL) is available free-of-charge after five p.m. The lot is reached from Centennial Drive by turning onto the "Gauss Way" access road that leads to MSRI / SSL. Disabled parking spaces are located at the entrance to MSRI's Chern Hall. Additional parking is available in the Institute's terraced hillside loop of parking spaces (upper two lots), which is accessed directly off Centennial Drive and requires walking up 95+ steps to the MSRI building.

For directions to MSRI, please see the URL below:

http://www.msri.org/about/directions/index_html

BIOGRAPHICAL SKETCH OF DR. ODELL