Vendredi 25 Juin 2010 à 11h, Salle des Conseils de l’IPN, Bâtiment 100
Soutenance de thèse de Carlos Eduardo Alvarez Cabrera (co-tutelle : Université Paris-Sud 11 et Université Los Andes -Bogota, Colombie) :
On Shape and Electrostatics : Statistical Mechanics Studies of Model Systems
This doctoral work presents a series of studies of systems in which the particles interact
by
means of spheroidal hard core and electrostatic interactions. We first consider a dipolar
hard
sphere bilayer, studied by means of Monte Carlo (MC) simulations. The pressure between the
layers is found to vary as -1/h⁵ , where h is the distance between layers. We observed
vortex
like structures, frustrated by the finite size of the system.
Next we obtained the analytical solution to the screened potential of charged spheroidal
colloid particles in the Debye-Huckel regime for Neumann and Dirichlet boundary
conditions. This latter result agrees with the solution of the Poisson-Boltzmann equation
far
from the colloid for strongly charged particles. We also perform MC simulations of
spheroidal colloids with a point charge at the center and find, both analytically and in
simulations, that the effective potential is stronger in the direction where the curvature
of the
colloid is higher.
Finally we present two studies under progress. The first one deals with the effect of the
addition of small spheres to spheroidal prolate particles in the nematic phase. We have
seen
that for an aspect ratio of 3 this effect is mild, but for an aspect ratio of 4, the
nematic to
isotropic transition is shifted to higher spheroid densities. In the second work,
preliminary
MC results for a size bidisperse spherical charged colloidal system are provided. This
allows
for a test of recently proposed mean-field approaches for polydisperse charged systems
(cell
model and renormalized jellium). In addition, we have found that as the size difference
between the colloids is increased, the screening of the smaller species increases, while
the
opposite effect is observed for the larger species.
Jeudi 10 Juin 2010 à 14h, au LPTMS, Bâtiment 100, salle 201
Marco Cosentino Lagomarsino ( Génophysique - Univ. Pierre et Marie Curie )
: The constrained growth of functional and
evolutionary partitioning of a genome.
The many available sequenced and annotated genomes can be partitioned
in different classes related to gene function and evolution. These
partitioning sometimes displays remarkably collective behavior. For
example, genomes of the same size tend to produce the same number of
protein families, regardless of specific evolutionary or environmental
traits. Clearly, the trends bringing to this collective behavior must
transcend at least some of the specificity of a genome. Rather, they
must be the product of more "universal" aspects of genome growth.
I will discuss how scaling laws for two different genomic partitionings,
related to protein domain topologies and their functional annotations,
are related, and present minimal genome growth models that reproduce the
scaling laws, and can be analyzed using simple tools from statistical
mechanics.
Jeudi 3 Juin 2010 à 14h, au LPTMS, Bâtiment 100, salle 201
Petr Sulc ( Los Alamos National Laboratory )
: Belief propagation for graph partitioning
We study the belief propagation algorithm for the graph bi-partitioning problem, i.e. the
ground state of the ferromagnetic Ising model at a fixed magnetization. Application of a
message passing scheme to a model with a fixed global parameter is not banal and we show
that the magnetization can in fact be fixed in a local way within the belief propagation
equations. Our method provides the full phase diagram of the bi-partitioning problem on
random graphs, as well as an efficient heuristic solver that we anticipate to be useful in
a wide range of application of the partitioning problem.
Work done in collaboration with L. Zdeborova
Jeudi 27 Mai 2010 à 14h, au LPT, Bâtiment 210, salle 114
Camille Aron ( Laboratoire de Physique Théorique et Hautes Énergies, Université Pierre & Marie Curie, Paris ) : Driven Quantum Coarsening
We solve the dynamics of an ensemble of disordered quantum rotors
coupled to two leads at different chemical potential letting a current
flow through the system and driving it out of equilibrium.
We discuss the phase diagram in the temperature, voltage, strength of
quantum fluctuations and coupling to the bath. We uncover the
coarsening regime. In particular, the scaling functions of correlation
and response at late epochs are proven to coincide with their
classical counterparts.
Jeudi 6 Mai 2010 à 14h, au LPT, Bâtiment 210, salle 114
Royce Zia ( VirginiaTech, Department of Physics, USA ) : Convection cells without shear or gravity :
Surprises from a simple model in non-equilibrium statistical mechanics
A clear signature of far-from-equilibrium systems, convection cells
are ubiquitous in nature. Typically, they are driven by external
forces, like shear or gravity in combination with temperature
gradients. Here, we show the existence of such cells in a simple
(possibly the simplest) system involving only a temperature gradient.
In particular, we study a 2-D Ising lattice gas in contact with two
thermal reservoirs. For simplicity, we impose a sharply localized
temperature gradient. With the higher T maintained at "infinity" while
setting the lower T below the Onsager temperature, convection cells
emerge. Other motivations, further surprises, and some theoretical considerations, will also be presented.
ATTENTION : JOUR EXCEPTIONNEL
Mardi 13 avril 2010, 14h, Salle 114, Bâtiment 210
Peter Forrester (University of Melbourne) : A finale for spacing distribution in random matrix theory
Recently a numerical analyst Folkmar Bornemann has shown that
the most efficient way to compute spacing distributions in classical
random matrix theory is to use Fredholm determinant formulas. I will
review these developments, including some contributions which come from
my own work.
Jeudi 25 Mars 2010 à 14h, au LPTMS, Bâtiment 100, salle 201
David Lacoste (Laboratoire PCT, ESPCI, Paris) :
Relations de fluctuations pour les moteurs moléculaires
Les moteurs moléculaires sont des molécules capables
de convertir l’énergie chimique issue de l’hydrolyse de l’ATP en
travail mécanique. Ces systèmes opèrent
typiquement loin de l’équilibre, dans un
régime dans lequel les approches thermodynamiques classiques
ne s’appliquent pas. Sur des modèles simples de moteurs moléculaires,
issus des modèles de type ratchet, on présentera certains aspects
de cette conversion d’énergie, et on discutera l’existence de
relations de fluctuations très générales, valables arbitrairement loin
de l’équilibre comme la relation de Gallavotti-Cohen en particulier.
Vendredi 19 Février 2010 à 14h, au LPTMS, Bâtiment 100, salle 201
ATTENTION : JOUR INHABITUEL
Ivan Junier (Institut des Systemes Complexes Paris Île-de-France (ISC-PIF)
rue Lhomond 75005 Paris, France ) :
Insights into genetic regulation : confronting polymer physics
and genomics
There is increasing evidence that the overall regulation of genetic
expression relies on a proper spatio-temporal organization of
chromosomes inside the cells. This can be seen in particular from the
transcription machinery : in some eukaryotes and bacteria,
transcription of highly active (co-functional or co-regulated) genes
occurs within discrete foci called transcription factories, where RNA
polymerases, transcription factors and their target genes co-localize.
I will show that the spatial coordination of co-functional genes
relies on two complementary types of genomic organization whose use
depends on the number of co-functional genes to coordinate. Then,
using single chain polymer theory, I will present a possible mechanism
for the formation of transcription factories, which allows to
understand the importance of genomic organization in the process.
In a second part, I will show that DNA inter-gene distances can be
used to efficiently work out functional links between genes. In
particular, I will present a network — the closome — that reflects
the tendency for COGs (cluster of orthologous genes) to be close to
each other along the DNA. It has been obtained by analyzing the
genomic organization of more than 800 different bacteria.
Jeudi 25 Février 2010 à 14h, au LPTMS, Bâtiment 100, salle 201
Kirsten Martens (Laboratoire de Physique de la Matière Condensée et des Nanostructures,
Université Lyon 1) : Dependence of the Fluctuation-Dissipation Temperature on the Choice of
Observable
On general grounds, a nonequilibrium temperature can be consistently
defined from generalized fluctuation-dissipation relations only if it is
independent of the observable considered. In this talk I will argue that
the dependence on the choice of observable generically occurs when the
phase-space probability distribution is nonuniform on constant energy
shells. I will relate quantitatively this observable dependence to a
fundamental characteristics of nonequilibrium systems, namely, the
Shannon entropy difference with respect to the equilibrium state with
the same energy. The results will be illustrated on various mean-field
models.
Jeudi 18 Février à 14h, au LPT, Bâtiment 210, salle 114
Paolo Pierobon (Inserm U932, Institut Curie, Paris) : Intracellular traffic models and single molecule beyond the membranes
Intracellular transport along microtubules or actin filaments, powered
by molecular motors such as kinesins, dyneins or myosins, has been
recently modeled using one-dimensional driven lattice gases. I will
discuss some generalizations of these models, that include extended
particles and defects. I will also present the most recent experiments
on single molecule study inside the live cell, where I measured the
velocity, processivity and step size of Myosin V. Finally I will
attempt a connection between theory and experiments.
Jeudi 21 Janvier à 14h, au LPT, Bâtiment 210, salle 114
Laetitia Gauvin (Laboratoire de Physique Statistique, ENS, Paris) : Analysis of socio-economic sytems from a statistical physics point of view : case of Schelling segregation model
The Schelling’s segregation model introduced in the 1970’s has attracted much attention in
social sciences and more recently in statistical physics. Here, we study the collective behavior of
this socio-economic model making used of methods borrowed from statistical physics following an
approach whose relevance was not conspicuous.
The main model parameter is the "tolerance" which quantities how much an agent is willing
to accept neighbors with different socio-economic characteristics. The dynamics of the model may
lead to segregated states, i.e. to large domains composed of agents with similar characteristics. To
study this dynamics through numerical simulations, we introduce a measure of segregation based
on percolation theory. We argue that this measure and the tolerance parameter play roles similar
respectively to the order parameter and the temperature in spin models. Based on this analogy,
we introduce the associated specific heat and susceptibility. This physical approach allows to
distinguish quantitatively several regimes and to characterize the transitions between them, leading
to the building of a phase diagram. Some of the transitions evoke empirical sudden ethnic turnovers.
We exhibit a dynamically frozen phase reminiscent of glassy or kinetically constrained systems. We
also establish links with spin1 models in physics, in particular the Blume-Emery-Griffiths and
Blume-Capel models for binary mixtures and alloys in the presence of vacancies. Although the
model studied can be considered as a toy model in comparison with the real social systems, our
approach provides generic tools to analyze the dynamics of more complex socio-economic systems.
Jeudi 14 Janvier 2010, 14h00, Salle 114, Bâtiment 210
Konrad Hinsen (Centre de Biophysique Moléculaire, Orléans) : Les cristaux de protéines : un pont entre les mondes microscopiques et macroscopiques
Un grand nombre de cristaux de protéines sont fabriqués tous les jours dans les laboratoires de recherche dans le but d’observer la structure atomique de leurs constituants par cristallographie. Mais ces cristaux peuvent aussi être étudiés à l’échelle macroscopique en tant que matériau. Pour faire le lien entre ces deux échelles, il est nécessaire d’avoir un modèle mécanique du cristal entier à l’échelle microscopique. Comme premier pas dans le développement d’un tel modèle, je montrerai comment un modèle simple pour la dynamique des protéines (le modèle du réseau élastique) peut être étendu à un cristal entier. Il permet de calculer des quantités microscopiques (les fluctuations des positions atomiques) et macroscopiques (la vitesse de propagation du son dans le cristal) accessibles toutes les deux à l’expérience.
Jeudi 17 Décembre 2009, 14h00, Salle 114, Bâtiment 210
Raphael Chetrite (Physics of Complex Systems, Weizmann Institute of Science) : Fluctuations d’un système markovien : du théorème de fluctuation-dissipation au niveau 2,5 des grandes déviations.
Je présenterai durant ce séminaire divers éléments de la physique d’un système Markovien hors
d’équilibre. Dans une première partie, je montrerai comment obtenir une cinématique, une vitesse
typique de ces processus dont la trajectoire est en général non différentiable.
Je montrerai alors que dans le référentiel Lagrangien associé à cette vitesse typique, le système
devient un processus d’équilibre si le système initial était un processus markovien sans saut.
Dans une seconde partie, je présenterai d’abord un rapide survol de l’ensemble des relations de
fluctuations, puis je me restreindrai au cas d’une dynamique homogène en temps (stationnaire).
Je montrerai alors que la relation de Gallavotti-Cohen qui est une symétrie de la fonction de grande
déviation de la production d’entropie (niveau 1 des grandes déviations = pour les observables) peut
être généralisée en une symétrie du niveau 2,5 des grandes déviations (pour le temps local et les
densités empiriques qui décrivent une trajectoire).
Jeudi 10 Décembre à 14h, au LPTMS, Bâtiment 100, salle 201
Herve Isambert (Institut Curie, Centre de Recherche, CNRS-UMR168, Paris) : Evolution of biomolecular networks under duplication-divergence processes
Genomic duplication-divergence processes are the primary source of new protein
functions and thereby contribute to the evolutionary expansion of functional
molecular networks. Yet, it is still unclear to what extent such
duplication-divergence processes also restrict by construction the emerging
properties of molecular networks, regardless of any specific cellular functions.
We have addressed this question motivated by the evolution of protein-protein
interaction networks, transcription networks and signal transduction networks.
Our approach relies on a general duplication-divergence model, based on the
statistically necessary deletions of physical or regulatory interactions,
arising from stochastic duplications at various genomic scales, from single
gene to whole genome duplications.
Major evolutionary scenarios are shown to depend on two global parameters
only : i) a protein conservation index (M), that controls the evolutionary
history of biomolecular networks, and ii) a distinct topology index (M’)
controlling their resulting structure. We then demonstrate that conserved,
non-dense networks, which are of prime biological relevance, are also
necessary scale-free by construction, irrespective of any evolutionary
variations or fluctuations of the model parameters. This turns out to result
from a fundamental linkage between individual protein conservation and network
topology under general duplication-divergence evolution.
By contrast, the conservation of network motifs including at least one
interaction cannot be indefinitely preserved under general
duplication-divergence evolution (independently from any network rewiring
dynamics), in broad agreement with empirical evidences between
phylogenetically distant species. This affects, in particular, the resolution
of regulatory conflicts in duplicated transcription networks, as well as, the
decoupling of signaling pathways in duplicated signal transduction networks.
All in all, these evolutionary constraints, inherent to duplication-divergence
processes, appear to have largely controlled the overall topology and
scale-dependent conservation of large biomolecular networks.
Jeudi 26 Novembre à 14h, au LPT, Bâtiment 210, salle 114
Nicolas Dupuis (Laboratoire de Physique Théorique de la Matière Condensée,
Université Pierre et Marie Curie, Paris) : Comportement infrarouge et fonction spectrale de bosons en interaction
Nous montrons que le groupe de renormalisation non-perturbatif (NPRG) apporte une solution simple au problème des divergences infrarouges qui apparaissent dans la théorie de la suprafluidité lorsqu’on pousse l’approche perturbative au-delà de la théorie de Bogoliubov. Le NPRG montre qu’en dimension d\leq 3, la théorie de Bogoliubov n’est plus valable pour des échelles de longueur supérieures à une longueur caractéristique 1/k_G, elle-même supérieure à la longueur de "réparation" (healing length). Dans le régime infrarouge (p\ll k_G) , le comportement de la fonction de Green à une particule est entièrement déterminé par les symétries (identités de Ward) et la divergence de la susceptibilité longitudinale induite par le couplage entre fluctuations transverses et longitudinales. Le mode de Bogoliubov (mode de Goldstone de la phase suprafluide) coexiste avec un continuum d’excitations, en accord avec les prédictions de l’approche hydrodynamique de Popov. Nous soulignons le lien entre ces résultats et le comportement critique à la transition suprafluide-isolant de Mott observée expérimentalement dans les gaz bosoniques ultrafroids en présence d’un réseau optique.
Jeudi 19 Novembre à 14h, au LPTMS, Bâtiment 100, salle 201
Thierry Mora (Lewis-Sigler Institute for Integrative Genomics, Princeton University) : Statistical physics modeling of the bacterial flagellar motor
Many bacteria like E. coli swim by virtue of small rotary motors that
drive rotation of helical flagella. Each motor is powered by a
transmembrane proton flux passing through the motor. This flux is
converted into torque with near-perfect efficiency by a mechanism
whose details remain largely unknown. First I will describe the
important biophysical properties of the motor, as measured in
experiments, including the recent observation of a stepping behaviour
at low speeds. I will then present a simple physical model that allows
us to explain most of these data, but also to make new predictions. In
particular, I will show how steps can be interpreted as
barrier-crossing events in a corrugated energy landscape. Then I will
show how to use our model to calculate the effect of shot noise (due
to the discrete nature of the energy source—the protons) on motor
diffusivity, and thus propose experiments to measure the proton
cooperativity in the torque generation process.
Vendredi 23 Octobre 2009, 14h00, Salle 114, Bâtiment 210
Lenka Zdeborova (Los Alamos National Laboratory, U. S. A.) : Long time dynamics in mean field glassy models
We introduce a generalization of the cavity method that
allows to follow Gibbs states in mean field systems when an external
parameter, e.g. the temperature, is tuned. Our method yields a static
way to study long time dynamics in mean field models of glass formers,
spin glasses, or random constraint satisfaction problems.
For the optimization problems the method gives limits of performance
for the simulated annealing algorithm. It also gives new results on
stability of the non-equilibrium states, the spinodal point upon
heating, we also demonstrate the presence of temperature chaos, and
study consequences of the shape of the energy landscape for
algorithmic hardness. Reference : Krzakala, Zdeborova ;
arXiv:0909.3820v1.