Introductory talk: Biological programming

• Jeremy J. Ramsden (University of Basel)

Room: Amphi LABRI

The Origin of Individuals

• Jean-Jacques Kupiec (Centre Cavaillès, Ecole normale supérieure, Paris)

Room: Amphi LABRI L’évolution des espèces (phylogenèse) et le développement des organismes individuels (ontogenèse) sont considérés comme deux phénomènes distincts. La biologie repose sur cette séparation qui pose l’espèce et l’individu comme principes premiers, réels et coextensifs, l’espèce étant une collection d’individus identiques. Dans sa version moderne cette ontologie s’appuie sur la théorie du programme génétique : une espèce est une collection d’individus possédant le même programme génétique et l’évolution des espèces est le résultat des mutations qui affectent leurs programmes (théorie synthétique de l’évolution). Cette conception est aujourd’hui mise en question par les données expérimentales. En effet, la théorie du programme génétique repose sur l’idée que les interactions des molécules biologiques excluent l’alea et qu’elles sont spécifiques. Au contraire, les données récentes montrent que les protéines manquent de spécificité. Elles peuvent interagir avec de nombreuses molécules partenaires. En conséquence, les interactions moléculaires sont intrinsèquement probabilistes, y compris dans la chromatine et l’expression des gènes est également un phénomène probabiliste. Cela contredit la théorie du programme génétique à sa racine. La prise en considération du manque de spécificité des protéines et du caractère intrinsèquement probabiliste des interactions entre molécules biologiques induit une nouvelle conception. La sélection naturelle agit non seulement dans la phylogenèse mais aussi l’ontogenèse. Celle-ci, au lieu d’être un processus déterministe dans lequel l’information génétique circule uniquement des gènes vers le phénotype (l’organisme individuel), est au contraire probabiliste et duale : les gènes fournissent les protéines, mais leurs interactions probabilistes sont triées par les contraintes sélectives produites par les structures cellulaires (et multicellulaires), qui sont elles mêmes soumises à la sélection naturelle. Au final, cette conception débouche elle-même sur une nouvelle ontologie : il n’existe qu’un seul phénomène d’ontophylogenèse expliqué par la seule théorie de sélection naturelle agissant en même temps sur l’ontogenèse et la phylogenèse. J.J. Kupiec, L’origine des individus, Fayard, 2008. (The Origin of Individuals, World Scientific, 2009).

A Reactive Multiagent Program using Stream Processing to Simulate Multicellular Systems.

• Pascal Ballet (Université de Bretagne Occidentale)

Room: Amphi LABRI Programming multiagent systems for biological systems close to cellular scales require significant computational capabilities for different reasons: the studied systems include many interacting entities (usually more than 10^5), they have very different sizes (from 10^-8 for the macromolecules to 10^-4 for small organisms) and their behaviors are frequently complex. For example a cell changes its shape, migrates, orientates, divides, adheres, communicate by direct contact or at distance with others cells, etc. The speed of microprocessors, in term of clock frequency, reaching its limits (around 4GHz), major manufacturers have chosen to develop multicore processors. This evolution is seen in Central Processor Units (CPU) but also in Graphic Processor Units (GPU). GPU contains stream processors allowing the parallel treatment of data (Single Instruction Multiple Data). Moreover, with the incoming of the openCL language, it is now possible to use he power of stream processing on different hardware (mother board and graphic cards for instance). Thanks to this, we have developed reactive multiagent algorithms and a simple software architecture to simulate multicellular phenomena with many interacting entities (more than 10^5). We show different simulations like a simple random walk, a Belousov-Zhabotinsky like reaction, a prey-predator system and a multicellular morphogenesis system. The examples are detailed and their efficiencies and drawbacks are discussed.

Collective discussion on MAS simulation frameworks Room: Amphi LABRI

Stochasticity in gene expression, evolution and evolvability

• Jean-Pascal Capp (LISBP - INSA de Toulouse)

Room: Amphi LABRI

Development of the Cabbage Root fly in agricultural landscape : seeing agroecosystems from a fly point of view.

• Nicolas Parisey (INRA, Rennes)

Room: Amphi LABRI

A Stochastic Model of cellular differentiation

• Bertrand Laforge (LPNHE)

Room: Amphi LABRI

A multi-scale agent-based model for the simulation of avascular tumor growth

• Guillaume Hutzler (Evry-Val d'Essonne University)

Room: Amphi LABRI

Exploring Hierarchical Evolution with an Artificial Protocell

• Barry McMullin (Rince Research Institute, Dublin City University)

Room: Amphi LABRI

Hydrodynamic simulation of ventral mesoderm invagination during Drosophila Melanogaster gastrulation

• Philippe-Alexandre Pouille (Instituto de Biología Molecular de Barcelona)

Room: Amphi LABRI The mechanical aspects of embryonic morphogenesis have been in most cases simulated using finite element models, which describe the tissue as a continuous medium. Here we develop a simulation of Drosophila embryo invagination of its ventral mesoderm during gastrulation, that allows access to both cellular and multicellular mechanical behaviours of the embryo. This model can be viewed as multi-agent where the individuals are the cell membranes characterized by an acto-myosin cortical tension and connected by apical and basal junctions and an acto-myosin contractile ring at the apical junctions. They interact with each other through hydrodynamic flow. Behaviours observed in vivo, including apical junction movements at the onset of gastrulation, cell elongation and subsequent shortening during invagination, and the development of a dorso-ventral gradient of thickness of the embryo, are predicted by this model as passive mechanical consequences of the genetically and biochemically controlled increase in the apical surface tension in invaginating mesoderm cells. In a second step, we also implemented the biochemical control system we investigated through experiments on the embryo. Here, a second set of individual agents are the cells and their gene expressions. We showed that ventral invagination initiation can be explained by a positive mechanical feedback. Under this hypothesis, the simulations account for the phenotypes observed in wild-type embryos and all the main mutants for invagination.

A new 'hierarchical dynamic networks' approach to multi-scale structure-function modeling of the kidney

• Randy Thomas (IBISC, FRE 2873 CNRS/Université d'Evry)

Room: Amphi LABRI

Estimation of model parameters: application to the complex I of the respiratory chain

• Christine Nazaret (IMB, Universite Bordeaux 1)

Room: Amphi LABRI Developing dynamic models of metabolic pathways of the whole mitochondria could help in predictive or preventive medecine. But theses pathways are necessarily large and complex with lots of variables that can not be measured directly in vivo currently. Besides the different existing rates that describe the kinetic behavior of enzymes in this network are mostly non linear (in terms of parameters and variables) which leads to difficulties in adjusting the parameters. That is why searching for a simple rate that describe the kinetic behavior of enzymes is useful. In this context, we derive an equation, called EMA (Extended Mass Action) Next we studied our abilities to adjust the parameters of this new equation theoretically and numerically. Finally we apply it to our experimental data derived from complex 1 enzymatic reaction.

Biological simulations : focus on 3D environment with multi-agent systems

• Marie Beurton-Aimar (LaBRI, Université Bordeaux)

Room: Amphi LABRI

Multi-agent System based on social agents: a GPU application

• Richard Moussa (LaBRI, Université Bordeaux)

Room: Amphi LABRI

Unsupervised learning to assist modeling of multilevel complex systems

• Javier Gil-Quijano (Laboratoire d'Informatique,Paris VI)

Room: Amphi LABRI The modelling by simulation of complex systems is a cyclic process: the modeller incorporates his/her knowledge into the model, runs simulations, discovers bugs or unwanted effects, corrects the model and eventually his/her knowledge, and the cycle restarts. The process ends when it is not possible to further improve the model because of technical or knowledge limitations. That cyclic process is particularly hard when modelling multi-level complex phenomena mainly because of the emergence of high level structures : the behaviour of lower level agents can be strongly influenced by the existence of emergent structures. Those structures must then be detected and considered as agents in simulations. The detection of structures is particularly difficult because of their dynamic nature. To consider the structures as agents implies to provide them with a behaviour. The latter is not an easy task because of the interdependence of the behaviours of agents that are placed at different levels. In this talk I propose to adress those problems by using automated learning mechanisms. In the first part of this talk I propose the use of statistical learning to discover the emergent structures. In the second part I propose the use of technics of automatic composition of programs to build the agents’ behaviours.

Towards a Multi-Level Modeling Language to Represent and Specify Emergent Structures in Agent-Based Model

• VO Duc An (UPMC Univ Paris 6)

Room: Amphi LABRI

Mobile Agents for Large Distributed Computing Systems

• Mohamed Mosbah (LaBRI, Université Bordeaux)

Room: Amphi LABRI The concept of mobile agents has been introduced recently as a novel and powerful paradigm to facilitate the design and programming of distributed applications. However, while their popularity continues to grow, a uniform theory of mobile agent systems is not yet sufficiently elaborated, in comparison with classical models of distributed computation. In this talk, I present a model based on local computations to encode mobile agent algorithms. In doing so, we approach a general and unified framework for expressing mobile agent computations which is consistent with the classical theory of distributed algorithms based on local computations.