Search

Discrete modelling frameworks of Biological networks can be divided in two distinct categories: Boolean and Multi-valued. Although Multi-valued networks are more expressive for qualifying the regulatory behaviours modelled by more than two values, the ability to automatically convert them to Boolean network with an equivalent behaviour breaks down the fundamental borders between the two approaches. Theoretically investigating the conversion process provides relevant insights into bridging the gap between them. Basically, the conversion aims at finding a Boolean network bisimulating a Multi-valued one. In this article, we investigate the bisimilar conversion where the Boolean integer coding is a parameter that can be freely modified. Based on this analysis, we define a computational method automatically inferring a bisimilar Boolean network from a given Multi-valued one.; Les cadres de modélisation discrets des réseaux biologiques se distinguent en deux catégories: booléennes et à valeurs multiples. Bien que les réseaux à valeurs multiples soient plus expressifs pour qualifier le comportement des réseaux moléculaires, modélisés par plus de deux valeurs, la possibilité de les convertir automatiquement en réseau booléen avec un comportement équivalent amenuise les frontières fondamentales entre ces deux approches. Une étude théorique du processus de conversion permet de combler l'écart entre eux. Fondamentalement, la conversion a pour enjeu de trouver un réseau booléen bisimulant un réseau à valeurs multiples. Dans cet article, nous étudions la conversion bisimilaire où le codage entier booléen en est un paramètre qui peut être librement modifié.De cette analyse, nous définirons une méthode de calcul inférant automatiquement un réseau booléen bisimilaire à partir d'un réseau à valeurs multiples donné.

International audience; Complex diseases such as Cancer or Alzheimer's are caused by multiple molecular perturbations leading to pathological cellular behavior. However, the identification of disease-induced molecular perturbations and subsequent development of efficient therapies are challenged by the complexity of the genotype-phenotype relationship. Accordingly, a key issue is to develop frameworks relating molecular perturbations and drug effects to their consequences on cellular phenotypes. Such framework would aim at identifying the sets of causal molecular factors leading to phenotypic reprogramming. In this article, we propose a theoretical framework, called Boolean Control Networks, where disease-induced molecular perturbations and drug actions are seen as topological perturbations/actions on molecular networks leading to cell phenotype reprogramming. We present a new method using abductive reasoning principles inferring the minimal causal topological actions leading to an expected behavior at stable state. Then, we compare different implementations of the algorithm and finally, show a proof-of-concept of the approach on a model of network regulating the proliferation/apoptosis switch in breast cancer by automatically discovering driver genes and their synthetic lethal drug target partner.; Les maladies complexes telles que le cancer ou la maladie d'Alzheimer sont causées par de multiples perturbations moléculaires conduisant à un comportement cellulaire pathologique.L'identification des perturbations moléculaires induites par la maladie et le développement de thérapies efficaces dépendent fondamentalement de la complexité de la relation génotype-phénotype. En conséquence, une question clé est de développer des cadres reliant les perturbations moléculaires et les effets des médicaments à leurs conséquences sur les phénotypes cellulaires. De ce fait un enjeu majeur consisterait à identifier les ensembles de facteurs moléculaires causaux menant à une reprogrammation du phénotype. Dans cet article, nous proposons un cadre théorique appelé réseaux de contrôle booléens, dans lequel les perturbations moléculaires et les actions médicamenteuses induites par une maladie sont perçues comme desactions topologiques sur des réseaux moléculaires conduisant à une reprogrammation du phénotype cellulaire. Nous présentons une nouvelle méthode utilisant les principes de raisonnement abductif déduisant les actions topologiques causales minimales conduisant au comportement attendu à l'état stable. Ensuite, nous proposons différentes implémentations de l’algorithme et montrons enfin une preuve de concept de l’approche sur un modèle de réseau régulant le commutateur prolifération / apoptose dans le cancer du sein en découvrant automatiquement les gènes conducteurs et leur partenaire cible létal synthétique.

In this article, we propose a domain speci c language, GUBS (Genomic Uni ed Behavior Speci cation), dedicated to the behavioral speci cation of synthetic biological devices, viewed as discrete open dynamical systems. gubs is a rule-based declarative language. By contrast to a closed system, a program is always a partial description of the behavior of the system. The semantics of the language accounts the existence of some hidden non-speci ed actions possibly altering the behavior of the programmed device. The compilation framework follows a scheme similar to automatic theorem proving, aiming at improving synthetic biological design safety.

This paper proposes a new framework for cooperative games based on mathematical relations. Here cooperation is defined as a supportive partnerships represented by a directed network between players (aka hedonic relation). We examine in a specific context, modeled by abstract games how a change of supports induces a modification of strategic interactions between players. Two levels of description are considered: the first one describes the support network formation whereas the second one models the strategic interactions between players. Both are described in a unified formalism, namely CP~game. Stability conditions are stated, emphasizing the connection between these two levels. We also stress the interaction between updates of supports and their impact on the evolution of the context.

International audience; A high-level Petri net framework is introduced for the toxic risk assessment in biological and bio-synthetic systems. Unlike empirical techniques mostly used in toxicology or toxicogenomics, we propose a systemic approach consisting of a series of behavioral rules (reactions) that depend on abstract discrete "expression" levels of involved agents (species). We introduce a finite state high-level Petri net model allowing exhaustive verification (model-checking) of properties related to equilibrium alteration or appearing of hazardous behaviors. The approach is applied to the study of the impact of the aspartame assimilation into the blood glucose regulation process.

International audience; Cancer is a complex and dynamic process caused by a cellular dysfunction leading to a whole organ or even organism vital perturbation. To better understand this process we need to study each of the levels involved, what allows the scale change, and to integrate this knowledge. A matricellular protein, PAI-1, is able to induce in vitro cell behaviour modifications, morphological changes, and to promote cell migration. PAI-1 influences the mesenchymoamaeboid transition. This matricellular protein should be considered as a potential launcher of the metastatic process acting at the molecular, cellular, tissular levels and, as a consequence, at the organism's level.

In this paper, we present a games network based modeling applied to Plasminogen Activation system (PAs). Games network theory extends game theory by including the locality of interactions. Each game in a games network represents a set of local interactions between some biological agents. We first introduce main features of the model, then we describe the modeling of the Plasminogen Activation system which is a signal transduction pathway involved in cancer cell migration. The PAs system is implicated in cytoskeleton modifications via actin and microtubules regulations, which in turn favours cell migration. Finally, we briefly present an extension of this work which consists on designing a simplified model (game) from an initial one. The method is based on abstract interpretation and aims at reducing the capabilities of each agent (strategies) while preserving its behaviour.

Automatic gene annotation is a basic requirement to identify the precise location of all the genes in a genome. Today however, the highest quality gene annotations remain those generated by human experts based on different reliable resources. Here we describe Exogean, a new automatic gene annotation method that combines two such resources, i.e. cross-species sequence alignments and mRNA sequences, using the same heuristic rules as human experts. Exogean is based on directed acyclic coloured multigraphs (DACMs) where nodes represent biological objects (exons, transcripts), and multiple edges between nodes represent relations between objects derived from human expertise. DACMs are read along paths that replicate the rules applied by human annotators when processing data. By representing relations and rules separately, the use of DACMs establishes a strict independence between the data and the heuristics applied to the data. Exogean is thus modular and flexible, and represents a generic framework that may capture any human expertise in the field of eukaryotic gene annotation. We describe the evaluation of Exogean on human chromosome 22 in comparison to eight other gene annotation methods, and we evaluate the influence of the evolutionary distance and the exhaustiveness of the input resources on these performances. The Exogean source code in OCaml is available on request at djebali@ens.fr.

International audience; In this paper we present an extension of game theory named games network. The objective of this extension is to propose a theoretical framework which suits to ``modular dynamics'' resulting from different local interactions between agents. Briefly, games networks describe the situation where each player plays different games at the same time with several different players. The theoretical extension is based on strategic games. We more particularly focus on the determination of a global equilibrium from local Nash equilibria. Especially we determine some conditions on the existence of a global mixed equilibrium whatever could be the representation of given interactions by a game network. The results are based on an intermediary representation called the dependence graph which describes interactions between players.