1,721,170 research outputs found
A Generalized rapid development environment for cellular automata based simulations
No full textCellular Automata (CA) are widely applied in variety of fields, and this made generalised simulation environments become increasingly important for the development of CA-based scientific applications. In this paper we discuss the fact that many real phenomena require strong relaxation of classical CA assumptions in order to be adequately modelled, and that brings about limitations of existing modelling and simulation environments, often based on insufficiently generalised CA formulations. These considerations have induced us to develop a modelling environment based on a largely generalised CA formulation. The environment has proven to be particularly suitable for modelling and simulation of spatial urban, territorial and environmentally-oriented phenomena
Analysis of large biological data: metabolic network modularization and prediction of N-terminal acetylation
No full textDuring last decades, biotechnology advances allowed to gather a huge amount of biological data. This data ranges from genome composition to the chemical interactions occurring in the cell. Such huge amount of information requires the application of complex algorithms to reveal how they are organized in order to understand the underlying biology. The metabolism forms a class of very complex data and the graphs that represent it are composed of thousands of nodes and edges. In this thesis we propose an approach to modularize such networks to reveal their internal organization. We have analyzed red blood cells' networks corresponding to pathological states and the obtained in-silico results were corroborated by known in-vitro analysis. In the second part of the thesis we describe a learning method that analyzes thousands of sequences from the UniProt database to predict the N-alpha-terminal acetylation. This is done by automatically discovering discriminant motifs that are combined in a binary decision tree manner. Prediction performances on N-alpha-terminal acetylation are higher than the other published classifiers
Numerical Modelling of Confluent Cell Monolayers : Study of Tissue Mechanics and Morphogenesis
No full textIn this thesis, we present the numerical model of confluent cell monolayers that we developed in order to study the interplay between cell biophysical properties and tissue mechanics and morphology. Our cell based model combines both cell physics and biology; it includes the representation of cell mechanics, the application of external constraints, as well as the simulation of cell proliferation and signalling. This model was inspired from the vertex model of Farhadifar et al. 2007, which we adapted and progressively extended within our EpiCells framework. Our model allowed us to investigate (i) how cell mechanical properties affect the response of tissues to stretching and how cells may adapt to the resulting strain, (ii) the development of the spine follicles covering the lower back of Acomys Dimidiatus, and (iii) tissue buckling. Finally, we extended our 2D model to the 3D space and presented its application for further studies of tissue folding
Dynamics on complex networks: application to abstract micro-economical models
No full textThis thesis presents a new class of methods and tools dealing with complex systems and their dynamics. We investigate the emerging behaviours of three stylized micro-economical models that simulate goods, labour, and credit markets, the three main ingredients that sustain an economy. In a goods market, economic actors exchange goods against cash. In labour market, economic actors exchange working hours against cash. In credit market, economic actors lend and redeem cash between them. We found that the structure of interactions between components plays a key role on the emerging behaviours. A cyclic interaction stabilizes goods and labour markets, but accumulates debt on credit market and drives to systemic avalanche of defaults. Some stylized facts in real economy are reproduced such as the Zipf law and the exponential growth of total accumulated debts. In controllability point of view, we found that acyclic interactions need power-law control and cyclic interactions need exponential control
Parameter Estimation of Platelets Deposition: Approximate Bayesian Computation With High Performance Computing.
Get PDFCardio/cerebrovascular diseases (CVD) have become one of the major health issue in our societies. Recent studies show the existing clinical tests to detect CVD are ineffectual as they do not consider different stages of platelet activation or the molecular dynamics involved in platelet interactions. Further they are also incapable to consider inter-individual variability. A physical description of platelets deposition was introduced recently in Chopard et al. (2017), by integrating fundamental understandings of how platelets interact in a numerical model, parameterized by five parameters. These parameters specify the deposition process and are relevant for a biomedical understanding of the phenomena. One of the main intuition is that these parameters are precisely the information needed for a pathological test identifying CVD captured and that they capture the inter-individual variability. Following this intuition, here we devise a Bayesian inferential scheme for estimation of these parameters, using experimental observations, at different time intervals, on the average size of the aggregation clusters, their number per mm2, the number of platelets, and the ones activated per μℓ still in suspension. As the likelihood function of the numerical model is intractable due to the complex stochastic nature of the model, we use a likelihood-free inference scheme approximate Bayesian computation (ABC) to calibrate the parameters in a data-driven manner. As ABC requires the generation of many pseudo-data by expensive simulation runs, we use a high performance computing (HPC) framework for ABC to make the inference possible for this model. We consider a collective dataset of seven volunteers and use this inference scheme to get an approximate posterior distribution and the Bayes estimate of these five parameters. The mean posterior prediction of platelet deposition pattern matches the experimental dataset closely with a tight posterior prediction error margin, justifying our main intuition and providing a methodology to infer these parameters given patient data. The present approach can be used to build a new generation of personalized platelet functionality tests for CVD detection, using numerical modeling of platelet deposition, Bayesian uncertainty quantification, and High performance computing.SCOPUS: ar.jinfo:eu-repo/semantics/publishe
Modelling and analysis of nonlinear distributed parameters systems using the Lattice Boltzmann method : application to free surface shallow water
No full textNous étudions dans cette thèse, composée de deux parties, la modélisation des écoulements en eaux peu profondes par la méthode de Boltzmann sur réseau et l'analyse des propriétés de commandabilité et d'observabilité des modèles obtenus. Dans la première partie, nous nous consacrons d'abord à la modélisation par la méthode de Boltzmann sur réseau des équations de Saint-Venant. En utilisant une linéarisation autour d'un profil d'équilibre, une représentation sous forme d'état des modèles de Boltzmann sur réseau est définie. Cette représentation incorpore les termes de force, et permet une définition complète des entrées (commandes) et des sorties (mesures). Nous représentons ensuite les phénomènes de sédimentation dans les écoulements en eaux peu profondes avec la méthode de Boltzmann sur réseau. Ce modèle défini en une dimension est validé numériquement en le comparant avec un modèle de volumes finis qui résout les équations de Saint-Venant-Exner. Le modèle LB défini est moins gourmand en temps de calcul et plus facile à manipuler que les modèles traditionnels. Dans la deuxième partie, nous traitons de l'analyse des propriétés de commandabilit é et d'observabilité des modèles LB obtenus. La première analyse est faite sur les critères algébriques de Kalmann et permet d'établir la non conservation des propriétés de commandabilité et d'observabilité lorsque l'ordre de réduction du système est augmenté. Une analyse plus approfondie basée sur la détermination les grammiens de commandabilit é et d'observabilité montre également que le constat reste valide pour les méthodes de discrétisation classique. La résolution des grammiens est faite avec des méthodes particulièrement adaptées aux structures creuses et de grande dimension que sont les matrices de la dynamique, de commande et/ou d'observation des modèles LB. Enfin, nous établissons que pour une commande aux frontières classique des canaux d'irrigation en débit et hauteur, la famille de systèmes des modèles LB d'ordre réduit n'est pas uniformément commandable alors qu'avec l'utilisation des variables de scattering comme variables de commande, cette famille devient uniformément commandable.We study in this thesis, subdivided into two parts, the modeling of the free surface shallow water flows with the lattice Boltzmann method and the analysis of the properties of controllability and observability of the resulting models. The first part focusses on the modeling of the shallow water flows with the lattice Boltzmann method. Using a linearization around a given equilibrium profile, we give a state space representation of the defined lattice Boltzmann models. This representation takes into account the force term, and allows a complete definition of the inputs (controls) and outputs (measures) variables. After this, we extend the model to include the phenomena of sedimentation. The defined one-dimensional model is validated numerically by comparing it with a finite volume model which solves the Saint-Venant-Exner's equations. The defined LB model is less complex (from a numerical point of view) and easier to handle. In the second part, we deal with the analysis of the properties of controllability and observability of the models obtained from the LB modeling of the shallow water flows. The first analysis, which is done with the Kalmann's algebraic criterias, leads to the establishment of the loss of controllability when the number of discretization sites increases. An extensive analysis, based on the determination of the controllability and observability gramians, allows to show that this conclusion remains with the classical methods of discretization. The determination of the gramians is done with particular methods well suited for the sparse and large matrices which are the dynamical, control and/or observation matrices of the LB models. Finally, we establish that for a classical boundary control of the irrigation canal with flow and level, the family of LB systems variables is not uniformly controllable, while using scattering variables as the control variables, the family becomes uniformly controllable
Optimizing the behavior of a moving creature in software and in hardware
No full textWe have investigated a problem where the goal is to find automatically the best rule for a cell in the cellular automata model. The cells are either of type OBSTACLE, EMPTY or CREATURE. Only CREATURE can move around in the cell space in one changeable direction and can perform four actions: if the path to the next cell is blocked turn left or right, if the path is free, i. e. the neighbor cell is of type EMPTY: move ahead and simultaneously turn left or right. The task of the creature is to cross all empty cells with a minimum number of steps.
The behavior was modeled using a variable state machine represented by a state table. Input to the state table is the neighbors state in front of its moving direction.
All combinations of the state table which do not expect a trivial or bad behavior were considered in software and in hardware in order to find out the best behavior. The best four-state algorithm allows the creature to cross 97 % empty cells on average over the given initial configurations.
As software simulation, optimization and learning methods are very time consuming, parallel hardware is a promising solution. We described this model in Verilog HDL. A hardware synthesizing tool transforms the description into a configuration file which was loaded into a field programmable gate array (FPGA). Hardware implementation offers a significant speed up of many thousands compared to software
Understanding signal sequences with machine learning
No full textProteins synthesized in the cell must be transported to the correct cellular compartment so that they can achieve their function. This process is a fundamental aspect of cell protein metabolism. All the proteins that must be secreted, carry a particular region of conserved function, the signal sequence (SS) or signal peptide, located in N-terminal extremity. To address the problem of correctly discriminating secreted proteins from the other ones (cytosolic), artificial intelligence techniques have been considered. The training set was composed of E. coli proteins whose location was determined experimentally. We used a set of wild type proteins completed by two mutants sets: (i) 15 SS which have lost their function and (ii) 240 proteins which gained SS function. We used evolutionary computing to generate new features able to better predict secretion. The idea here was to extend existent theory. To reach this goal, we designed a generic framework to described physico-chemical requirements. To reduce the huge number of amino-acid properties to a tractable amount, we proposed a clustering method using on a novel correlation based distance. Resulting performances are higher than preceding attempts on wild-type proteins (95.8% of cross-validated accuracy), but also on mutant collections. Furthermore, the new properties give new insights about the signal sequence requirements. An analysis of these new features, along with their usage in the decision trees, allowed us to explain some apparently contradicting experiments about signal sequence secondary structure
- …
