April 21 - 25, 2008
The goal of my research program is to better understand how biological circuits program multicellular phenotypes. While a major focus of the group involves phenotypes associated with human epithelial systems, we have also drawn upon more tractable model systems, such as C. elegans and yeast, to glean deeper fundamental insights. Each of these three biological systems provides unique advantages that my lab has sought to exploit using a combination of computational and quantitative experimental approaches.
Networks of biological signals guide cells to form
multicellular patterns and structures. Understanding the
design and function of these complex networks is a fundamental
challenge in developmental biology and has clear implications
for biomedical applications, such as tissue engineering and
regenerative medicine. Signaling networks are composed of
highly interconnected pathways involving numerous molecular
components. Precisely to what extent multicellular structures
are susceptible to quantitative variations in underlying
signals and to what extent Nature has utilized this mechanism
of "quantitative diversification" during evolution are unclear.
I will describe a computational framework that we have
developed to explore and to quantify the multicellular
diversity that emerges from signaling perturbations. We have
applied this method to study vulval development in C.
The approach is not only effective in predicting the molecular
genetics of multicellular patterning, but also gauges the
capacity of this signaling network for creating phenotypic
diversity. In fact, model predictions strongly correlate to
multicellular phenotypes observed across ten species related to
C. elegans. These results suggest that systems-level modeling
can shed insight into the evolutionary trajectories of
regulatory networks that gave rise to divergent multicellular
We attempt to address the question of why there are so many intermediate species in biochemical
reaction networks using an idea from realization theory. We describe how prescribed biological
function can be designed with very low dimensional models, which are however not implementable
with the physically allowable biochemical reaction mechanisms. It then becomes apparent that
the introduction of a large number of intermediate species can be interpreted as a realization technique
to enable the implementation of prescribed function with the available dynamical building blocks.
By reversing this realization scheme, we propose a model reduction paradigm for biochemical reaction networks.
Human language has long captured the imagination of biological
researchers, but the gulf separating 'computation,' 'biology', and
'language' has been equally long-standing: the classical biological
problem of how to bridge between a genotype and a phenotype,
in this case, perhaps the most complex behavioral phenotype
we know of. The aim of this talk is to show how recent developments
in linguistic theory bridge this 'abstraction gap' by
illuminating some of the modular design properties of human language,
illustrating that despite its apparent surface complexity, human language's
core seems to be far simpler than has been previously supposed,
potentially reducible to a single, simple, basic operation that derives
all the seemingly special properties of human language. Further, by
positing this modular approach, we can shed light on the computational
interfaces of human language to the systems of speech/motor production and parsing,
as well as internal systems of inference; here there seems to be
a natural, and expected kind of 'impedance matching,' with the design
seemingly forced to follow constraints imposed by considerations of
semantic interpretation, rather than considerations of computational
complexity in parsing or production. Finally, this new modular view leads naturally to
evolutionary considerations as to how these components arose in the course of
evolution, in this case, a more 'saltational' view than is generally supposed.
Joint work with Qiang Zhang(1), Melvin E. Andersen(1)
and Rory B.
The terminal differentiation of B cells in lymphoid organs into
antibody-secreting plasma cells upon antigen stimulation is a
step in the humoral immune response. The architecture of the
transcriptional regulatory network consists of coupled
mutually-repressive feedback loops involving the three
factors Bcl6, Blimp1 and Pax5. This structure forms the basis
irreversible bistable switch directing the B-cell to plasma
differentiation process - i.e., the switch remains on even
activating stimulus (antigen) is removed. The environmental
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) is known to suppress
humoral immune response by interfering with this
program. We have developed a computational model of the
pathways that regulate B-cell differentiation, and the
mechanism by which TCDD impairs this process through the action
aryl hydrocarbon receptor (AhR). Using a kinetic model and
analysis, we propose that TCDD regulates the proportion of
differentiate into plasma cells by raising the threshold dose
lipopolysaccharide (LPS) required to trigger the
We also show that stochastic modeling of gene expression, which
cell-to-cell differences in content of signaling proteins,
distributional characteristics to the timing and probability of
differentiation among a population of B-cells. This
variability is likely to be a key determinant of dose-response
sensitivity of individual cells to differentiation.
(1) Division of Computational Biology, The Hamner Institutes
Sciences, Research Triangle Park, NC 27709, USA
National Center for Computational Toxicology, U.S.
Protection Agency, Research Triangle Park, NC 27711, USA
Joint work with BW Kim.
The Phenotype Genotype Project on Addiction and Depression (PGP), with support from the Departments of Neurology, Psychiatry, and Radiology, and Center for Human Genetic Research; Massachusetts General Hospital and Harvard Medical School; Boston, MA, USA
Positive and negative preferences can be assessed by keypress procedures that quantify (i) decision-making regarding approach, avoidance, indifference, and ambivalence responses, and (ii) judgments that determine the magnitude of approach and avoidance. Most prior studies of relative preference have not used keypress procedures, but have used ratings of personal utility, as a composite index of approach and avoidance. Such composite ratings can be calibrated against an absolute framework such as the macroeconomic pricing of commodities, as is done with Prospect Theory. It leaves open questions of whether or not splitting composite ratings of utility into approach and avoidance measures reveals any patterns in behavior, as observed for Prospect Theory. Such patterns might include a trade-off relationship between approach and avoidance, or a value function that might not need an absolute macroeconomic framework. We assessed these questions across multiple sets of healthy control sub
jects with a keypress procedure, and found a set of patterns for approach and avoidance are (i) recurrent across many stimulus types, and (ii) robust to the injection of noise. These patterns include: (a) a preference trade-off plot that counterbalances approach and avoidance responses and represents biases in preference and consistency/uncertainty of preference, (b) a value function linking preference intensity to uncertainty about preference, and (c) a preference saturation plot that represents how avoidance actions are over-determined relative to approach actions. One can consider this set of patterns to be a form of relative preference theory (RPT), since they meet the same requirements for recurrency and robustness to noise as the Weber-Fechner-Stevens Law. As with the value function for Prospect Theory, the value function in RPT has a steeper slope for negative relative to positive preferences (i.e., loss aversion), and can be described as a power law, or a logarithmic functi
on. All of these patterns show symmetry between group and individual data in that they have similar mathematical formulations as manifolds or boundary asymptotes for group data, or as fitted functions for individual data across multiple variables. These patterns verify known biases between females and males regarding viewing beautiful and average faces. When used to evaluate cocaine dependent subjects versus healthy controls, these patterns quantify the phenotype of the restricted behavioral repertoire observed in addiction. When used as regressors in the analysis of fMRI data, RPT measures are associated with significant BOLD signal change across a set of reward/aversion brain regions. Both keypress behavior and fMRI BOLD signal can be further associated with polymorphisms is genes such as CREB1 and BDNF. Given RPT scaling between groups of subjects and individuals, further work is warranted to assess if scaling can be achieved to brain circuitry and genetics.
The human brain is a network containing a hundred billion neurons, each communicating with several thousand others. As the wiring for neuronal communication draws on limited space and energy resources, evolution had to optimize their use. This principle of minimizing wiring costs explains many features of brain architecture, including placement and shape of many neurons. However, the shape of some neurons and their synaptic properties remained unexplained. This led us to the principle of maximization of brain’s ability to store information, which can be expressed as maximization of entropy. Combination of the two principles, analogous to the minimization of free energy in statistical physics, provides a systematic view of brain architecture, necessary to explain brain function.
The concept of function has a tainted history in science and philosophy, having been mated to teleology and harnessed to pull various discredited theories. This presentation begins with a short history designed to give a sense of what made the concept of function problematic. Current philosophical attention to the problem revolves ellipse-like around two foci: systematic or system relative accounts, and selectionist accounts. These are briefly described, and the systematic account is defended on the grounds that it captures a ubiquitous explanatory strategy in science and engineering, and that selectionist accounts are, in contrast, limited to evolutionary biology (and, perhaps, designed artifacts), and make it difficult to articulate various important issues in evolutionary biology itself.
Joint work with Xiling Shen, Justine Collier, Lucy Shapiro, Mark Horowitz, and Harley H. McAdams.
We developed a mechanistic model of the cell cycle control of Caulobacter Crescentus. Symbolic model checking reveals that the cell cycle is extremely robust to parameter variations, and that the cell cycle starts and stops reliably to accommodate arbitrary starvation periods.
Determining proper sensitivity to incoming signals is important to all regulated biological systems, but is especially crucial for responses that involve an irreversible decision. Using Xenopus oocytes maturation as an example, we illustrate how entangled feedback architectures can be used to integrate the multiple signals needed to make the switchlike, irreversible transition from interphase to meiosis. We also discuss how the specific topologies of these feedback loops modulate the switching threshold as a function of the progesterone input.
The evolution of communication provides one of the few
examples in evolutionary biology where principles of physical
acoustics, interacting with developmental constraints on
physiology and motor control, have clear and predictable
effects on evolutionary outcomes. I will provide two clear
example of this, relating to basic physiological constraints
on fundamental and formant frequencies, and various
morphological "tricks" that vertebrates have evolved to evade
such constraints. Then I will explore a more speculative
hypothesis concerning the neural control of vocalization:
that the production of complex learned vocalizations requires
particular neural correlates (direct fronto-bulbar
connections) and behaviours ("babbling" or subsong), and
furthermore that constraints on development may mean that
there are only a few ways to achieve such vocal capabilities,
raising the interesting possibility of "deep homology" in the
evolution of communication. This hypothesis, though
speculative, is testable and is consistent with the most
recent information on genes involved in vocal learning.
Many gene regulatory networks are modeled at the mesoscopic scale, where
chemical populations change according to a discrete state (jump)
Markov process. The chemical master equation for such a process is
typically infinite dimensional and unlikely to be computationally
tractable without reduction. The recently proposed Finite State
Projection technique allows for a bulk reduction of the CME while
explicitly keeping track of its own approximation error. We show how a
projection approach can be used to directly determine the
statistical distributions for stochastic gene switch rates, escape times,
trajectory periods, and trajectory bifurcations, and to evaluate how likely
it is that a network will exhibit certain behaviors during certain intervals
of time. We illustrate these ideas through the analysis of the switching
behavior of a stochastic model of Gardner's genetic toggle switch.
I will review the control of localization by rat vibrissa system.
Understanding the effects of blood viscosity
a very crucial role in hemodynamics, thrombosis and
could provide useful information for diagnostics and therapy of
vascular disease. Blood viscosity, which arises from frictional
interactions between all major blood constituents, i.e. plasma,
proteins and red blood cells, constitutes blood inherent
flow in the blood vessel. Because red blood cells (RBCs) are
constituent of the cellular phase of blood, white blood cells
platelets normally do not have a great influence on whole blood
viscosity. When blood flows through a vessel, it tends to
two different phases. In direct contact with the wall a low
phase exists, which is deficient in cells and rich in plasma
and acts as
a lubricant for the blood transport. In the central core region
vessel a high viscosity phase exists, which depends on the
In this paper, the nature and stability of blood flow in a
is investigated numerically using a spectral collocation
expansions in Chebyshev polynomials. The study reveals that a
hematocrit concentration in the central core region of a large
has a stabilizing effect on the flow.
Keywords: Arterial blood flow; Hematocrit; Variable viscosity;
stability; Chebyshev spectral collocation technique
1.) Pedley T. J.: The Fluid Mechanics of Large Blood Vessels.
University Press, London, 1980.
2.) Makinde O. D.: Magneto-Hydromagnetic Stability of plane-
flow using Multi-Deck asymptotic technique. Mathematical &
Modelling Vol. 37, No. 3-4, 251-259, 2003.
3.) Makinde O. D. and Mhone P. Y.: Temporal stability of small
disturbances in MHD Jeffery-Hamel flows. Computers and
Applications, Vol. 53, 128–136, 2007.
4.) Makinde O. D.: Entropy-generation analysis for
channel flow with non-uniform wall temperature. Applied Energy,
Here, we present a transport model which indroduce a variant in
the transport theory. Here the mass or merchandize in order to
go from a port to a destination must pass through a deposit,
without accumulation. It is possible to consider deposit
constraint. We obtain the corresponding linear program
associated to it, as well as the dual. The application of this
model to biological sciences, specially in human body, lungs, blood and biological systems seems promising.
We obtain the dual, which appear in a natural way from the
incidence matrix, which has many interesting properties.
The rank is the number of the ports, deposits plus the
destination minus on.
We study and characterize all the extremals extending the ideas
of Jurkar-Ryser, and von Neumann in the case of the classical
transpot model. A relation with the material in the papers by
E. Marchi:Z.Wahrscheinlichtheorie verww. geb,12,220-230
(1969) and 23,7-17, (1972), is underway.
We were succesful to extend this theory to mixed model
considering that in a deposit, there exists the possibility to
stay or passing. This rich study permits more potentiality to
the tools. An extension to several steps has been performed and
its potentiality for applications is vast.
Lotka-Volterra equations are very famous by two biological
spices. Generally the most common presentations are in ecology
science. We study this models without considering the
antisymmetric condition among the parameters.(see Marchi and
Velasco Revista Mexicana de Fisica,36,No4(1990)665-679).We
obtain a movement constant for the systems which satisfies the
Hamilton equation. For more than two interacting species the
result the Volterra result are biologically improper. This is
due to the antisymetric conditions. Following Volterra original
methods combined with the variation of Montroll et. all, who
applied such equation to maser and laser, we obtain some
powerful personal procedure that we applied to the three and
four species obtaining and general condition and build the
cycle. Moreover this can be extended for an arbitrary number of
biological species obtaining constructively the cycle. The
background idea is to delete one variable at that time using
partial differentials equations of second power of the first
order. Furthermore we consider the conservation of the density
of point in the fase space (LIoville theorem) an we postulate
that the same distributions following the Gibbs Canonical Law.
By the way Prigogyne at all applied Lotka-Volterra system in
non-equilibrium thermodynamics. Moreover, there are several
applications of Lotka-Volterra to the theory of membrane. We
can show you some of them. Finally, we point out that
there are about four hundred papers in the subject in the last
ten years, and non of them apply our methods. The potentiality
for real application is very important even if the system of
Lotka-Volterra is asymptotically unstable.
This talk will present an engineering perspective on "architecture" in complex engineered systems. The role of protocols and interfaces will be emphasized, along with other architectural concepts such as modularity, evolvability and reusability. Examples of architecture as applied to autonomous vehicles will be used to illustrate modern engineering design tools.
This talk will present an introduction to some of the key principles and tools from feedback control theory. The two main design principles that will be explored is the role of feedback as a tool for managing uncertainty, and the use of feedback to design the dynamics of a system. Examples from engineering and nature will be used to illustrate some of the key concepts and techniques.
Recently, numerous engineers have demonstrated that genetic circuits can
be effectively modeled and analyzed utilizing methods originally
developed for electrical circuits leading to new understanding of their
behavior. If this is possible, then it may also be possible to design
synthetic genetic circuits that behave like particular electrical
circuits such as switches, oscillators, and communication networks.
Synthetic genetic circuits have the potential to help us better
understand how microorganisms function, produce drugs more economically,
metabolize toxic chemicals, and even modify bacteria to hunt and kill
tumors. There are, however, numerous challenges to design with genetic
material. For example, genetic circuits are composed of very noisy
components making their behavior more asynchronous, analog, and
non-deterministic in nature. Therefore, design methods must be adapted
to consider these issues. Interestingly, future electrical circuits may
soon also face these challenges which opens up the very intriguing idea
that these new design methods may in the future be utilized to produce
more robust and power efficient electrical circuits.
Cellular information processing is carried out by complex biomolecular networks
that are able to function reliably despite environmental noise and genetic mutations.
The robustness and evolvability of biological systems is supported in part by neutral
networks and neutral spaces that allow for the preservation of phenotype despite underlying
genotypic variation. This talk will describe two such spaces. The first are sequence niches
that emerge in the process of satisfying constraints needed to avoid crosstalk among sets
of promiscuous and paralogous proteins. The second are highly anistropic, sloppy parameter
spaces that arise in multiparameter models of biomolecular networks. Implications of
the structure of these spaces for cellular information processing will be discussed.
Applying results from dynamical systems, control and optimization, we
develop new approaches for designing experiments to elucidate the
biochemical network structure of the chemotaxis mechanism in R.
sphaeroides. Biological information and data is used to create initial
models (model determination); an experiment is then designed in order
to discriminate between these models; and a model invalidation
procedure closes the loop. This way we can develop an understanding of
the underlying biochemical network structure and appreciate the
properties and architecture of bacterial sensory systems in general. A
Synthetic Biology approach can then be used to redesign such networks
for improved or modified functionality.
For the simple reproductive behavior exhibited by female quadrupeds, the neural circuit and several genomic modules have been worked out (Drive, MIT Press, 1999). Feedback is most prominent in the hormonal mechanisms that support the behavior, whereas most of the behavioral dynamics exhibit feedforward relations. Emboldened by the success of the analysis of this simple behavior, we've enlarged the focus of our lab's work to encompass sexual arousal and generalized CNS arousal (Brain Arousal, Harvard Univ. Press, 2006). Shannon's information theoretical calculations probably apply to CNS arousal and its (universal) complementary phenomenon, habituation. We have speculated (BioEssays, August 2007) that in an animal or human at rest, CNS arousal systems live in a chaotic domain and quickly go through a phase transition to orderly dynamics as the animal or human responds. We are testing this theoretical idea with simulations of neural nets and with deep brain stimulation that raises the arousal level of a brain damaged animal.
Determining quality of performance for a biological system is critical to identifying and elucidation its design principles. This important task is greatly facilitated by enumeration of regions within the system's design space that exhibit qualitatively distinct function. First, I will review a few examples of design spaces that have proved useful in revealing design principles for elementary gene circuits. Then I will present a recently developed approach to the generic construction of design spaces, illustrate its application to common classes of biochemical network motifs, and test predictions for a specific class with experimental data from human erythrocytes.
There is no question that "systematic" functional analysis, which explains the capacities or features of a system based on interactions among its parts, plays an important role in biology. But an important role is also played by "selectionist" functional analysis, where engineering or design principles carry explanatory weight based on the action of natural selection. While some have questioned the legitimacy of selectionist functional analysis, asserting that it involves illegitimate teleological notions or an incorrect understanding of evolutionary theory, careful work by philosophers of biology over the last 35 years has helped explicate and defend this approach. Uses of the concept of function in biology may be interpreted, based on context, in terms of one or the other type of functional analysis, and sometimes in terms of both, and thus pluralism about function should be embraced.
We formulate a mathematical version of the conventional model
of maintenance of silencing in S. cerevisiae and analyze the conditions
for bistability as well as for formation of stationary boundaries.
Although the model is perhaps too simple, the structure of the
bifurcation diagram, describing parameter regions with different kinds
of qualitative behavior, is likely to be more robust. We can place some
of the known mutants in different regions of this diagram. One
interesting finding of this study is that, under some conditions, the
lowering of acetylation rates might have non-obvious consequences for
silencing. Possible improvements of the model as well as experimental
tests are discussed at the end.
This talk present the basic ideas underlying a method developed by the speaker
and his collaborators, for the analysis of the dynamics of certain biomolecular
Measuring locomotor behavior of Drosophila in enclosed arenas is a powerful way to obtain a quantitative behavioral phenotype. The measures previously used for this purpose are relatively coarse, thus ignoring fine scale movements of the fly. More importantly, they do not explicitly take into account how the arena shapes the dynamics of the locomotion. We acquired data using a video-tracking system to measure the trajectory of a fly over extended periods as it explores a circular arena. Based on this data, we present some metrics that take into consideration the dynamics of the trajectory, as well as the interactions of the trajectory with the geometry of the arena. The basic idea is to treat the locomotor trajectory of the fly as a stochastic process, and then estimate a set of marginal distributions of the probability measure describing this process. The measures include joint or individual distributions of position, speed, path curvature, inter-event times and re-orientation angles after stopping. These probability measures can be worked into other behavioral setups, are relatively easy to calculate using robust statistical estimation procedures, and can account for environmental effects on the behavior. They also serve as foundation for a quantitative stochastic process model of the walking behavior.
It is well known that, for a molecule being produced in an unregulated manner and degraded exponentially (constitutive gene expression for example) the intrinsic noise in the molecule number will be such that variance equals the mean. This talk will establish fundamental limits on the amount of noise reduction that can be obtained by regulation when delays in the mechanism (due to transport and the finite time required to synthesize intermediate molecules) and when that mechanism has limited information carrying capacity in the sense of Shannon (due to itself involving finite numbers of molecules).
A number of Oscine songbirds are vocal learners. In these species, song shares a remarkable characteristic with human language: both are acquired through imitative learning. When reared in social and acoustic isolation, the Zebra Finch (the species of songbird under study here) can still sing, revealing the genetically encoded aspect of song. However, the isolate songs usually have longer syllables, and appear to be more variable than the wild type (WT) songs found in the wild or in laboratory colonies.
In order to understand how the isolate song might evolve over multiple generations, we successively trained naïve juveniles starting with an isolate founding father. The first generation learners are in turn used as tutors to train the next generation, and so on. Thus, tutoring lineages are established through a recursive training processes. We found that small, yet systematic variations accumulate over generations of training. Remarkably, the descendants’ song structures are gradually transformed towards WT songs.