Conference Dates: May 22 - 25, 2007
The Conference is devoted to speakers presenting their latest research in Network Science. The level of presentation is accessible to audiences outside of the speaker's respective field. In addition to keynote and invited talks we will have a number of contributed talks. An extended poster session will be held along with a "best poster" competition. Summarizing discussion sessions and panels will be also directed to facilitate common ground between diverse areas.
The talks may be downloaded in either ppt (if available) or pdf format. Click on the name of a speaker in the main table or just scroll down to find the speaker in the list. Download links are next to their schedule information. They are being posted, so please check back often for new postings.
Conference Program
TUESDAY 22 |
WEDNESDAY 23 |
THURSDAY 24 |
FRIDAY 25 |
||
| 7:30 - 8:15 | BREAKFAST |
||||
| 8:00-8:15 | OPENING |
||||
| 8:15-8:50 | |||||
| 8:50-9:25 | |||||
| 9:25-10:00 | |||||
| 10:00-10:15 | COFFEE BREAK |
||||
| 10:15-10:50 | |||||
| 10:50-11:25 | |||||
| 11:25-11:40 | |||||
| 11:40-11:55 | CLOSING |
||||
| 11:55-1:25 | LUNCH |
||||
| 1:25-2:00 | |||||
| 2:00-2:35 | |||||
| 2:35-3:10 | |||||
| 3:10-3:45 | POSTERS |
||||
| 3:45-4:00 | COFFEE BREAK |
||||
| 4:00-4:15 | POSTERS |
||||
| 4:15-4:30 | POSTERS |
||||
| 4:30-4:45 | |||||
| 4:45-5:00 | |||||
| 5:00-5:15 | Competition Award Ceremony | DiBONA-5:15 |
Panel: Networking Networks | ||
| 5:15-6:00 | Competition Award Ceremony | Open laptop session | Panel: Networking Networks | ||
| 6:15-7:15 | Keynote Talk: SAGAN |
||||
| 7:15-9:00 | Reception |
||||
| Legend: | Invited talks are 35 min = 30 min + 5 min questions Contributed talks are 15 minutes = 12 min + 3 min questions |
||||
|
Adamic, Lada (University of Michigan) Thu., 8:15 - 8:50 ppt pdf m4a m4v |
|
Title: |
Expertise Networks in Online Communities: Structure and Algorithms |
Abstract: |
Web-based communities have become an important place for
people to seek and share expertise. We find that networks in these
communities typically differ in their topology from other online networks
such as the World Wide Web. Systems targeted to augment web-based communities
by automatically identifying users with expertise, for example, need
to adapt to the underlying interaction dynamics. In this study, we analyze
the Java Forum, a large online help-seeking community, using social
network analysis methods. We test a set of network-based ranking algorithms,
including PageRank and HITS, on this large size social network in order
to identify users with high expertise. We then use simulations to identify
a small number of simple rules governing the question-answer dynamic
in the network. These simple rules not only replicate the structural
characteristics and algorithm performance on the empirically observed
Java Forum, but also allow us to evaluate how other algorithms may perform
in communities with different characteristics. We believe this approach
will be fruitful for practical algorithm design and implementation for
online expertise-sharing communities.
|
|
Albert, Réka (Penn State University) Tue., 9:25 - 10:00 m4a m4v |
|
Title: |
Topology-based logical modeling of biological network dynamics |
Abstract: |
Interaction networks between gene products form the
basis of essential processes like signal transduction or embryonic development;
cell-to-cell interactions determine organ function and pathogen-immune
system interactions. Recent experimental advances helped uncover the
structure of many molecular and cellular networks, creating a surge
of interest in the dynamical description of cellular or systemic regulation.
This presentation will explore the connections between network topology
and dynamics by introducing qualitative (logical) models of the signal
transduction network underlying plant responses to drought and of mammalian
immune responses to respiratory pathogens.
The first model uses a compilation of indirect experimental evidence to reconstruct and simulate the signal transduction process leading to closure of the microscopic pores on plant leaves. We find that the network is robust against a significant fraction of possible perturbations (gene disruptions or pharmacological interventions). The model offers a roadmap for the identification of candidate manipulations that have the best chance of conferring increased drought tolerance. The second model synthesizes experimental and clinical information on the interactions between host immune components and two closely related pathogenic bacteria in the genus bordetellae. Our results indicate that the infection time course of both bordetellae can be separated into three distinct phases based on the most active immune processes. The model offers predictions regarding cytokine regulation, key immune components and clearance of secondary infections; we experimentally validated two of these predictions. Although the biological systems and the technical details of the models differ, the models' success suggests a strong link between the topology and dynamics of biological networks. |
|
Almaas, Eivind (Lawrence Livermore National Lab) Tue., 2:00 - 2:35 ppt pdf m4a |
|
Title: |
Cellular metabolic network modeling |
Abstract: |
Network approaches have provided non-trivial insight on the organization of many biological
systems. In the case of cellular metabolism, where nodes correspond to metabolites and links
indicate chemical reactions, it is frequently possible to facilitate close interactions between
experiments and computational modeling. Using a "flux-balance" modeling approach, it is possible
to predict metabolic flux (or link weight) patterns on the organism-level, and study the effect
of network topology on cellular function and robustness. In this presentation, I will discuss
basics of metabolic network modeling and flux-balance calculations, using examples from the
whole-cell network reconstructions of metabolism in the bacterium Escherichia coli and the yeast
Saccharomyces cerevisiae. Our metabolic analysis has uncovered that, since the connectivity
distribution for all known metabolic networks is scale-free, the possible flux distributions are
characterized by a heavy-tail. Also, when subject to varying environmental conditions, persistent
metabolic network activity is centered on a core reaction set with special properties. Most
recently, we have used this modeling framework to study epistasis, the non-linear interaction between
genes as mediated by the metabolic network.
|
|
Bettencourt, Luis (Los Alamos National Laboratory) Fri. 8:50 - 9:25 pdf m4a |
|
Title: |
Quantifying Scientific Discovery: temporal evolution and network structure of six emerging fields. |
Abstract: |
It has long been argues that scientific discoveries generate new dynamics
and reorganizations of scientific communities.
We create temporal series and networks of co-authorship of six emerging
fields. The temporal dynamics of number of authors
is shown to be well described in terms of population models, which also permit forecasting
the future size of a field. Measures of scientific productivity are built by comparing the
increase in publication output of fields given their increase in numbers of authors and shown
to be well fit by scaling laws. The networks of co-authorship are built and analyzed
as the field develops.
|
|
Bonabeau, Eric (Icosystems, Cambridge, MA) Thu., 10:50 - 11:25 m4a m4v |
|
Title: |
A wealth of networks |
Abstract: |
Now that we have a nice emerging theoretical framework for understanding the
structure, dynamics and evolution of networks, we see networks everywhere.
As wonderful as this sounds, it begs the question: is it useful to see
networks everywhere? I will try to address this question from the
perspective of one of the possible usefulness currencies: money. Can the
emerging network science framework provide competitive advantage to
companies that know how to use it? I will focus in particular on social
networks as I believe this is where the question is most controversial.
Indeed, while it is clear that a better understanding of infrastructure and
physical networks can provide competitive insights into how to operate them
better, there is no such clarity when it comes to social networks. With the
help of several real commercial-world examples, I will try to shed some
light on this topic. Some of the conclusions may be surprising.
|
|
Börner, Katy (Indiana University) Tue., 2:35 - 3:10 pdf m4a m4v |
|
Title: |
Mapping the Evolving Interface of Mainstream Chemistry and the Fields of Biochemistry, Biology, and Bioengineering |
Abstract: |
How does our collective scholarly knowledge grow over time? What major areas of science
exist and how are they interlinked? Which areas are major knowledge producers; which
ones are consumers? Computational scientometrics - the application of bibliometric/ scientometric
methods to large-scale scholarly datasets - and the communication of results via maps of science
might help us answer these questions. This talk represents the results of a study that aims to
map the structure and evolution of chemistry research over a 30 year time frame. Information from
the combined Science (SCIE) and Social Science (SSCI) Citations Indexes from 2002 was used to
generate a disciplinary map of 7,227 journals and 671 journal clusters. Clusters relevant to
study the structure and evolution of chemistry were identified using JCR categories and were
further clustered into 14 disciplines. The changing scientific composition of these 14 disciplines
and their knowledge exchange via citation linkages was computed. Major changes on the dominance,
influence, and role of Chemistry, Biology, Biochemistry, and Bioengineering over these 30 years
are discussed. (This is collaborative work with Kevin W. Boyack and Richard Klavans.)
|
|
Bornholdt, Stefan (University of Bremen) Tue., 8:50 - 9:25 m4a m4v |
|
Title: |
Computation in molecular networks: Reliability despite biochemical stochasticity |
Abstract: |
How is living matter regulated so reliably, despite the molecular
and fluctuating nature of their central information processing circuits?
I will discuss this problem from two perspectives, in a toy model of
stochastic dynamical networks, and subsequently in the context of
biological examples. Mathematical toy models can teach us about the
conditions of when and how stochastic networks can self-synchronize into
a stable and reproducable dynamical pattern. The central question is how
a network of unreliable elements can function reliably, for example to
perform computations. This points us to possible construction principles
of real biochemical regulatory networks which we will discuss for some
known biological molecular networks.
|
|
Califano, Andrea (Columbia University) Mon., 8:15 - 8:50 ppt pdf m4a m4v |
|
Title: |
A human B lymphocyte interactome for the dissection of dysregulated pathways in lymphoid malignancies. |
Abstract: |
The identification of genes that are causally related to the presentation of
a specific malignant phenotype is still an open problem in cancer research.
The availability of a complete map of molecular interactions - including
transcriptional, complex-formation, metabolic, and signaling interaction -
would provide a rational basis for this research. Unfortunately, such a map
remains quite elusive. We have developed information theoretic methods to
predict both transcriptional (ARACNe) and post-translational (MINDY)
interactions in human B cells. Predictions from these methods have been
biochemically validated in vivo and shown to have very low false positive
rates. By combining these methods with other reverse-engineering algorithms
and high-throughput experimental data, using a standard Bayesian evidence
integration scheme, we have produced the first comprehensive draft of a human
B lymphocyte cellular network. We will discuss how such a draft can be used
to produce a map of interactions that are dysregulated in specific pathologic
or physiologic phenotypes. We also show how the dysregulation maps for three
specific B cell malignancies - including Mantle Cell, Burkitt, and Follicular
Lymphoma - can be used to pinpoint the causal lesions with high accuracy.
|
|
Cheswick, William R. (Lumeta Corporation) Tue., 1:25 - 2:00 pdf m4a |
|
Title: |
Mapping Internets and Intranets |
Abstract: |
The Internet Mapping Project started at Bell Labs in 1997. Hal Burch and Bill Cheswick it explored the
Internet daily using traceroute-style probes from a single host, and accumulated some 200 GB of trace data over
eight years. The project included attempts to visualize these large graphs using custom brute force layout
algorithms.
In 2000, Lumeta was spun off from the Labs to advance and apply this technology to intranets. Corporate and government networks are based on TCP/IP which, by design, resist central control and management. The maps and mapping tools allow improved control over large networks. But the examination of large graphs, and visualization of the differences and evolution of large graphs is an open research problem. A map of the full Internet is still very much a ball of yarn. |
|
Collins, James J. (Boston University) Wed., 8:50 - 9:25 m4a m4v |
|
Title: |
Engineering Gene Networks: Integrating Synthetic Biology & Systems Biology |
Abstract: |
Many fundamental cellular processes are governed by
genetic programs which employ protein-DNA interactions in regulating
function. Owing to recent technological advances, it is now possible
to design synthetic gene regulatory networks, and the stage is set for
the notion of engineered cellular control at the DNA level. Theoretically,
the biochemistry of the feedback loops associated with protein-DNA interactions
often leads to nonlinear equations, and the tools of nonlinear analysis
become invaluable. In this talk, we describe how techniques from nonlinear
dynamics and molecular biology can be utilized to model, design and
construct synthetic gene regulatory networks. We present examples in
which we integrate the development of a theoretical model with the construction
of an experimental system. We also discuss the implications of synthetic
gene networks for biotechnology, biomedicine and biocomputing. In addition,
we present integrated computational-experimental approaches that enable
construction of first-order quantitative models of gene-protein regulatory
networks using only steady-state expression measurements and no prior
information on the network structure or function. We discuss how the
reverse-engineered network models, coupled to experiments, can be used:
(1) to gain insight into the regulatory role of individual genes and
proteins in the network, (2) to identify the pathways and gene products
targeted by pharmaceutical compounds, and (3) to identify the genetic
mediators of different diseases.
|
|
D'Souza, Raissa (University of California, Davis) Fri., 8:15 - 8:50 pdf m4a |
|
Title: |
Optimization, Preferential Attachment, Viability and Network Growth |
Abstract: |
We show how the mechanism of preferential attachment can emerge from an
underlying network optimization framework. The preferential attachment
(PA) model so obtained has two novel features, saturation and viability,
which have natural interpretations in the underlying network. Like PA,
saturation has previously been assumed at an axiomatic level. The
combination of PA and saturation leads to power-law degree distributions
with exponential cutoff which give excellent fits to a broad range of
empirical observations of networks. Here we show how a simple underlying
optimization framework can give rise to both known mechanisms and likewise
to a new concept of viability, and suggest that such models form a good
starting point for the analysis of many networks. In addition we discuss
the fit provided to a broad range of data, including previously
unexplained data on the Internet obtained from "whois" tables.
|
|
Dunne, Jennifer A. (Santa Fe Institute) Wed., 8:15 - 8:50 ppt pdf m4a m4v |
|
Title: |
Ecological network structure, robustness, and uncertainty |
Abstract: |
It is increasingly apparent that an ecological network
perspective, which encompasses direct and indirect effects among interacting
taxa, is critical for understanding, forecasting, and managing the impacts
of species loss and invasion, habitat conversion, and climate change.
At a basic research level, this suggests that we need to develop a more
general framework for understanding ecological network robustness at
whole-system and component levels. Using examples of food webs from
the present, ~50 million years ago, and ~500 million years ago, I will
discuss how research at the interface of ecology and network theory
can be fruitfully extended across deep time, increasing our understanding
of different aspects of ecological robustness. I will also highlight
the importance of explicitly addressing uncertainty in data.
|
|
Gerstein, Mark (Yale University) Thu., 9:25 - 10:00 ppt pdf m4a m4v |
|
Title: |
Understanding Protein Function on a Genome-scale using Networks |
Abstract: |
My talk will be concerned with topics in proteomics, in particular
predicting protein function on a genomic scale. We approach this
through the prediction and analysis of biological networks, focusing
on protein-protein interaction and transcription-factor-target ones. I
will describe how these networks can be determined through integration
of many genomic features and how they can be analyzed in terms of
various simple topological statistics. In particular, I will discuss a
number of specific analyses: (1) Integrating gene expression data with
the regulatory network illuminates transient hubs; (2) Integration of
the protein interaction network with 3D molecular structures reveals
different types of hubs, depending on the number of interfaces
involved in interactions (one or many); (3) Analysis of betweenness in
biological networks reveals that this quantity is more strongly
correlated with essentially than degree; (4) Analysis of structure of
the regulatory network shows that it has a hierarchiel layout with the
"middle-managers" acting as information bottlenecks. (5) Development
of a useful web-based tools for the analysis of networks, TopNet and
tYNA.
|
|
Hausmann, Ricardo (Harvard University) Wed., 1:25 - 2:00 m4a |
|
Title: |
Macro consequences of a discontinuous product space |
Abstract: |
Much of economic theory has assumed that the product space is continuous
and smooth in the sense that there is always a product through which
countries can use their endowments and capabilities. This paper
summarizes recent work that has looked empirically at the patterns of
relatedness between products using network techniques and standard
economic tools. The main findings are that the product space is highly
heterogeneous with some dense sections surrounded by a sparse periphery.
The capacity of countries to move to higher end products depends
strongly on the existence of potential products that are near the areas
of current production. Since the product space is heterogeneous, this
explains several phenomena such as the differential growth rates of
countries, the lack of convergence of incomes at the global level and
the prolonged income collapses in developing countries.
|
|
Kleinberg, Jon (Cornell University) Wed., 2:00 - 2:35 pdf m4a |
|
Title: |
Anonymized Social Networks, Hidden Patterns, and Privacy Breaches |
Abstract: |
An increasing amount of social network research focuses
on large datasets obtained by measuring the interactions among individuals
who have strong expectations of privacy. To preserve privacy in such
instances, the datasets are typically anonymized -- the names are replaced
with meaningless unique identifiers, so that the network structure is
maintained while private information has been suppressed. I will discuss
recent joint work with Lars Backstrom and Cynthia Dwork in which we
identify some fundamental limitations on the power of network anonymization
to ensure privacy. In particular, we describe a family of attacks such
that even from a single anonymized copy of a social network, it is possible
for an adversary to learn whether edges exist or not between specific
targeted pairs of nodes.
|
|
Lazer, David M. J. (Harvard University) Thu., 1:25 - 2:00 ppt pdf m4a m4v |
|
Title: |
Life in the Network: The Coming Era of Computational Social Science |
Abstract: |
An increasing fraction of human behavior (especially relational behavior)
leaves substantial digital traces-- whether in the form of phone logs,
e-mail, instant messaging, etc. Further, increased computational power
allows the analysis of these digital traces-- e.g., through natural
language processing, statistical analysis of massive (millions of
individuals) longitudinal data, etc. These two points suggest that we are
on the precipice of dramatic new insights into collective human behavior. I
will discuss the potential future of a "computational social science", with
reference to four ongoing research projects.
|
|
Loscalzo, Joseph (Harvard University) Wed., 10:50 - 11:25 m4a m4v |
|
Title: |
Human Disease Classification in the Postgenomic Era: A Complex Systems Approach to Human Pathobiology |
Abstract: |
Contemporary classification of human disease dates
to the late 19th century, and derives from observational correlation
between pathological analysis and clinical syndromes. Characterizing
disease in this way established a schema that has served clinicians
well to the current time, relying on observational skills to define
syndromic phenotypes. Throughout the last century, this approach became
more objective as the molecular underpinnings of many disorders were
identified and definitive laboratory tests became an essential part
of the overall diagnostic paradigm. Yet, this classic diagnostic strategy
has widely recognized shortcomings that reflect both a lack of sensitivity
in identifying preclinical disease, and a lack of specificity in defining
disease unequivocally. In this presentation, I will focus on the latter
shortcoming, arguing that it is a reflection both of the different clinical
presentations of many diseases (variable phenotypic expression), and
of the excessive reliance on Cartesian reductionism in establishing
diagnoses. With advances in routine sequencing of the human genome,
evolving proteomic and metabolomic methodologies, and growing molecular
data sets from healthy and phenotypically well-characterized diseased
individuals, we are now in the unique position to consider all human
disease using rigorous, systems-based approaches. The purpose of this
presentation is to provide a logical argument for a new approach to
classifying human disease that both appreciates the uses and limits
of reductionism and incorporates the tenets of a non-reductionist complex
systems analysis. This approach offers the promise of diagnostic accuracy,
prognostic utility, and therapeutic efficacy that can serve as the objective
basis for the growing field of personalized medicine.
|
|
Makse, Hernán (Benjamin Levich Inst. and CCNY) Fri.,9:25 - 10:00 ppt m4a |
|
Title: |
Scaling, renormalization and self-similarity in complex networks |
Abstract: |
Our recent finding of scaling and topological self-similarity in
complex networks provides a new perspective on our view of biological
complexity. When we observe the networks of protein-protein
interactions and cellular metabolism with varying resolution, they
consistently show the self-replicating pattern of fractal with finite
fractal dimensions, which has a direct implication on the structural
stability and growth mechanism of the network. In particular, we show
the relevance of the scale transformation to the evolutionary process,
where the evolution of the cellular network follows the inverse of the
renormalization scheme. We investigate how the emerging topological
properties of biological networks were achieved in the long history of
evolution and how it is related with the error-tolerance level of the
network. We find that the scale-invariant topology implies a
significant increase in the robustness of the network, in accordance
with the established rules of natural selection.
We also characterize the large-scale modular organization in terms of
scale-invariant laws of modularity and develop a mathematical
framework to understand the emergence of the modular properties of the
network.
|
|
Maslov, Sergei (Brookhaven National Lab) Thu., 8:50 - 9:25 ppt pdf m4a m4v |
|
Title: |
Propagation of large concentration changes in reversible protein binding networks |
Abstract: |
We study how the dynamic equilibrium of the reversible
protein - protein binding network in yeast S. cerevisiae responds to
large changes in abundances of individual proteins. The magnitude of
shifts between free and bound concentrations of their immediate and
more distant neighbors in the network is influenced by such factors
as the network topology, the distribution of protein concentrations
among its nodes, and the average binding strength. Our primary conclusion
is that on average the effects of a perturbation are strongly localized
and exponentially decay with the network distance away from the perturbed
node. This explains why, despite globally connected topology, individual
functional modules in such networks are able to operate fairly independently.
We also found that under specific favorable conditions, realized fin
a significant number of paths in the yeast network, concentration perturbations
can selectively propagate over considerable network distances (up to
four steps). Such ”action-at-a-distance” requires high concentrations
of heterodimers along the path as well as low free (unbound) concentration
of intermediate proteins.
|
|
Masuda, Naoki (University of Tokyo) Thu., 2:35 - 3:10 pdf m4a m4v |
|
Title: |
Evolutionary games with participation costs on networks |
Abstract: |
Networks with heterogeneous degree distributions such
as scale-free networks are known to promote evolution of cooperation
in social dilemma games. With standard examples of payoff matrices,
cooperation is facilitated because hubs are more advantageous than others
by playing more often. I show that even a relatively small cost of participation
imposed on players neutralizes the constructive role of heterogeneous
networks in altruism. With participation costs, hubs are charged more
so that they do not spread cooperation. The participation cost is irrelevant
in many evolutionary dynamics on networks without degree dispersion
and in well-mixed populations, but it controls the level of cooperation
on heterogeneous networks.
|
|
Newman, Mark E. J. (University of Michigan) Wed., 9:25 - 10:00 pdf m4a m4v |
|
Title: |
Maximum likelihood methods for discovering structure in networks |
Abstract: |
Likelihood maximization is one of the fundamental tools
of data analysis in statistics but has yet to find widespread use in
the study of networks. This talk will describe two new methods for revealing
structure in networks using maximum likelihood methods. The first probes
for hierarchical structure in networks using a Markov Chain Monte Carlo
technique while the second attempts to classify network nodes into classes
according to their connection patterns using an expectation-maximization
algorithm. A variety of applications to real-world network data will
be described, including social, biological, and information networks.
|
|
North, Stephen (AT&T) Tue., 10:50 - 11:25 m4a |
|
Title: |
Scaling up network visualization |
Abstract: |
The past two decades of research have yielded surprisingly
effective methods for visualizing networks of hundreds of nodes, and
sometimes far larger ones. In this talk we will review progress to date
and discuss some recent work in the Graphviz project on visualizing
large networks. This includes better optimization for drawing real-world
networks, interactive adjustment of the level of detail within a network
viewer, and a method of extracting and displaying the subgraph "between"
two or more nodes in a quasi-random network (such as a social network,
possibly large enough to require external database storage).
|
|
Ravasz-Regan, Erzsébet (Harvard University) Thu., 2:00 - 2:35 pdf m4a m4v |
|
Title: |
Network Structure of Protein Folding Pathways |
Abstract: |
The classical approach to protein folding inspired
by statistical mechanics avoids the high dimensional structure of the
conformation space by using effective coordinates. Here we introduce
a network approach to capture the statistical properties of the structure
of conformation spaces, and reveal the correlations induced in the energy
landscape by the self-avoidance property of a polypeptide chain. We
show that the folding pathways along energy gradients organize themselves
into scale free networks, thus explaining previous observations made
via molecular dynamics simulations. We also show that these energy landscape
correlations are essential for recovering the observed connectivity
exponent, which belongs to a different universality class than that
of random energy models.
|
|
Roth, Fritz (Harvard University) Thu., 3:10 - 3:45 ppt pdf m4a m4v |
|
Title: |
MouseFunc 1: A Critical Assessment of Mammalian Gene Function Prediction Based on a Rich Evidence Network. |
Abstract: |
Several years after sequencing the human genome and
the mouse genome, much remains to be discovered about the functions
of most human and mouse genes. Computational prediction of gene function
promises to help focus limited experimental resources on the most likely
hypotheses. Several algorithms using diverse genomic data have been
applied to this task, but these have been primarily tested on the unicellular
yeast S. cerevisiae. We assembled a rich network of mouse functional
genomic evidence, such that each node is associated with a vector of
gene/protein properties, and each node pair is associated with a vector
of evidence for various biological relationships. Nine bioinformatics
teams used this rich network to independently train classifiers and
generate predictions of function for 21,603 mouse genes. We identified
strengths and weaknesses of current functional genomic datasets and
compared the performance of function prediction algorithms. This analysis
inferred functions for 76% of mouse genes, including five thousand currently
uncharacterized genes, with an average precision value of 35% at a recall
value of 20%.
|
|
Sagan, Dorion (Sciencewriters) Wed., 3:10-3:45 |
|
Title: |
Complex Systems in Energetic Context: How Real-World Complex Systems and Networks Accomplish Natural Goals |
Abstract: |
Although one of the simplest imaginable systems, the
heated air in Abe Lincoln's log cabin demonstrates
properties we associate with purpose, consciousness,
and planning. Wherever there is a crack or crevice in
the cabin, the heat acts as if it "wants" to escape.
Such behavior in simple close-to-thermal equilibrium
systems may lie at the root of the behavior of complex
living systems. In this lecture I provide a brief
survey of real--as opposed to algorithmic or
computer-generated--systems that appear in regions of
energy flow. The second law of thermodynamics can be
restated to say simply that energy spreads. As it
delocalizes gradients are reduced. The reduction of
ambient gradients in complex open systems appears to
be their raison d'etre. Arrogant humanity, and indeed
all evolving life, appear to belong to a general class
of naturally appearing complex systems. These systems
may lose parts of themselves and assemble into larger
systems that together more effectively accomplish the
mandate of energy dispersal than individuals. The
global economy, with its focus on energy acquisition,
and its disregard for human suffering and individual
freedoms, exemplify the inexorable second law no less
than heat escaping Lincoln's cabin.
|
|
Talas, Annamaria (Real Pictures) Fri., 10:50 - 11:25 |
|
Title: |
CONNECTING THE DOTS - The role of documentary films in public understanding of science |
Abstract: |
Communicating the new science of networks to a broad
television audience is a challenge. The author is an experienced documentary
filmmaker working for public broadcasters around the world. She has
been preparing a film for years on the origins and impact of network
science and will relate her experiences in communicating the scientific
concepts to broadcast commissioning editors, film funding agencies and
the general public in trying to raise interest in the project.
It has not been smooth sailing. Reactions range from disbelief to disinterest with occasional enthusiasm. While on the surface connectedness and networks seems all too obvious to the general public, on a deeper level they carry surprises that are not easy to comprehend. The science is still in its infancy and thus its implications are not yet widely understood. Without overwhelming practical applications it remains difficult to ‘sell’ a film based on concepts. Nevertheless, as the broad potential of network theory evolves, there is at last a growing acceptance of its potential public importance. |
|
Tang, Chao (University of California, San Francisco) Fri., 10:15 - 10:50 ppt pdf m4a m4v |
|
Title: |
Connecting function and topology of small biological circuits through dynamics |
Abstract: |
There is an intimate relationship between function and form (structure) on both
the macroscopic/organismic and the microscopic/molecular scales of the biological world.
On the "mesoscopic" scales, e.g. the cellular networks, this relationship is far less clear
and may have been masked by physical, biochemical, and evolutionary constraints. I will
demonstrate with a few examples that at the scale of small cellular circuits the requirement
of a robust function drastically limits the choice of network topologies for that function.
Such knowledge can help us to understand natural biological circuits as well as to synthesize
the artificial ones.
|
|
Uzzo, Stephen M. (The New York Hall of Science) Tue., 3:10 - 3:45 m4a m4v |
|
Title: |
Connections: The Nature of Networks A New Science Exhibition |
Abstract: |
The understanding of how networks function and evolve is important to the lives of
global citizens. It may provide a breakthrough in the ability of policy-makers,
teachers, voters, and public planners to grasp the otherwise elusive complex relationships
and behaviors in biological, ecological, social, and economic systems. "Connections: The
Nature of Networks," (NSF Award No. 0229268) is a unique public exhibition about networks
with an emphasis on characteristics of complex networks (social, biological, communications
networks, and others). It premiered in November, 2004 as part of a museum expansion project
sponsored by the City of New York. "Connections . . ." explores the fundamental structures
of networks (such as how nodes, links, and hubs form) and how they manifest themselves in what
visitors see around them in their daily lives, providing them with tools to understand
similarities and differences among various kinds of networks. Because complex networks have
the attribute of emergent behavior, the theme of emergence is an important aspect of the
exhibition. The talk will include an annotated tour of the exhibition and explore how it was
planned, executed, and evaluated; as well as some of the problems inherent in communicating
network science concepts to lay audiences. At the end of the tour there will be time for
questions and answer, as well as your comments on the exhibition.
|
|
Vespignani, Alessandro (Indiana University) Thu., 10:15 - 10:50 ppt pdf m4a m4v |
|
Title: |
The impact of mobility networks on the worldwide spread of epidemics |
Abstract: |
Networks which trace the activities and interactions
of individuals, transportation fluxes and population movements on a
local and global scale have been analyzed and found to exhibit large
scale heterogeneity, self-organization and other properties typical
of complex systems. Here we analyze the impact of mobility networks
on the global spreading of emerging infectious diseases. We define a
computational model for the large scale spread if infectious diseases
that integrates the air transportation network with demographic data.
The model is used to study the specific case of the SARS epidemic and
to provide scenario forecasts for pandemic influenza. The effect of
the network complexity on the predictability of the global spreading
pattern of emerging diseases is analyzed.
|
|
Vicsek, Tamás (Eötvös University, Budapest) Wed., 10:15 - 10:50 ppt pdf m4a |
|
Title: |
Statistics and evolution of the community structure of large social networks |
Abstract: |
Complex societal systems can be described in terms
of networks capturing the intricate web of connections among the units
they are made of. A fundamental question of great current interest is
how to interpret the global organisation of such networks as the coexistence
of their structural sub-units (called modules, communities, clusters,
etc) associated with more highly interconnected parts. Identifying these
unknown building blocks (e.g., industrial sectors, groups of people)
is crucial to the understanding of the structural and functional properties
of networks.
Here we first present an approach to analyse the main statistical features of the interwoven sets of overlapping communities, where two communities overlap if they have common members. Our studies of a variety of networks, including mobile phone call, collaboration, and school friendship graphs demonstrate that the web of modules has highly non-trivial correlations and specific scaling properties. Due to the versatility of our method we are able to follow the time development of individual communities and investigate quantitative aspects of social dynamics on a large scale. |
|
Vidal, Marc (Dana-Farber Cancer Institute) Wed., 6:15 - 7:15 |
|
Title: |
Interactome networks |
Abstract: |
For over half a century it has been conjectured that
macromolecules form complex networks of functionally interacting components,
and that the molecular mechanisms underlying most biological processes correspond
to particular steady states adopted by such cellular networks. However, until
recently, systems-level theoretical conjectures remained largely unappreciated,
mainly because of lack of supporting experimental data.
To generate the information necessary to eventually address how complex cellular networks relate to biology, we initiated, at the scale of the whole proteome, an integrated approach for modeling protein-protein interaction or "interactome" networks. Our main questions are: How are interactome networks organized at the scale of the whole cell? How can we uncover local and global features underlying this organization, and how are interactome networks modified in human disease, such as cancer? |
Title: |
Predicting evolution: a machine-learning approach to validating models for biological networks |
Abstract: |
The past decade has witnessed an explosion of interest and research activity
in statistical descriptions of real-world networks.
Early on, it was observed that certain select statistical attributes
of real-world networks could be reproduced by simple stochastic
models. This observation fueled the proliferation of
numerous such models -- particularly those purporting to describe
growing and evolving networks -- based on simple design principles.
However, as it became clear that multiple models could
reproduce varying attributes within a fidelity comparable to the
certainty with which they could be measured, it also became clear that
reproducibility did not imply predictability.
We present a resolution to this problem of over-universality -- the
problem of multiple design principles reproducing features observed in real-world networks -- by an
explicitly predictive approach. That is, by
exploiting machine learning approaches more commonly used
in high-dimensional statistical problems, we learn a prediction function,
which takes as input potentially hundreds or thousands of network attributes and
takes as output the `best' design principle: which of the several competing design principles is the most
consistent with the observed network. Our method is accurate (on held-out data
not seen while learning the prediction function) and interpretable (revealing {\it which}
topological attributes best distinguish the putative models).
We apply the method to several biological networks, using several machine learning algorithms,
with several choices of input attributes, and reveal a consistent trend: that
duplication mutation mechanisms are distinguishable from and
more consistent with real biological networks than preferential attachment mechanisms, and
that even similar duplication mutation design principles are topologically distinguishable.
|
Last Updated April 12, 2007 | Site Design by Elisha Hardy