|PDF Version||Biosystemics module modelling-engineering|
|Director : ROUX-ROUQUIE Magali (firstname.lastname@example.org)|
Our group is involved in the systemic modelling and simulation of biological systems and processes. Systemic modelling is the modelling of an entity as a reactive interacting component implementing processes over time (1-4). There can be several versions (discrete/continuous, deterministic/stochastic, synchronous/asynchronous, and so on) of the systemic metamodel and a large variety of formalisms and specification languages can be used . The expressiveness of the modelling languages depends on the underlying concepts, so the conceptual models must drive the theoretical choices and the objective ranking of the formalism relevance is better achieved using the conceptual models than using the methods applied to execute such models. Therefore, we have undertaken an extensive documentation of our conceptual models. This approach allowed us to assess the relevance of the object-oriented Unified Modelling language (UML) to systemic objectives and requirements: the UML allows the classification of (bio)entities according to static classes and roles, and also allows their behaviour to be addressed on a state-based description. Such behaviour is described in terms of transitions in-between states and these transitions are triggered by events and conditions Furthermore, UML formalism allows to account for the complexity of biological systems, in terms of nested levels (from molecules through supramolecular complexes, to cells) and quasi-decomposable systems. Each quasi-decomposable system enables us to differentiate a system into interacting subsystems, each module being modelled and interpreted in a relatively autonomous way. The number of subsystems depends on careful identification of the boundary relations between them. (5). Using a real-time extension of UML, UML-RT based on ROOM methodology (Rose Real-Time UML tools), we modelled and simulated MPF complex activation during the cell cycle. Molecules and molecular complexes were modelled in their cellular context as autonomous reactive software objects exhibiting parallel (concurrent) behaviour, using state-machines and collaboration diagrams. These state-machines were further compiled into timed-automata as used by the model checker (6). As the goal of UML is to gather several notations under a single language, additional assets are pertinent, and they include:
it supports high-levels of standardization, according to the Object Management Group (OMG) (for example, encoding UML in XML can be performed using the current OMG XMI standard),
it links tools for database access and formalized description and simulation of biological systems.
Our ongoing projects aim to improve understanding of the object-oriented modelling of the multifarious forms of biological entities according to their environment, and thereby to account for their multiple functions. In addition, we plan to pursue formalism ranking according to our systemic objectives and requirements by examining other approaches and tools (continuous languages, programming languages, Petri nets, etc.) currently used in systems biology. One of our middle-term goals is to investigate the co-design and the co-simulation of biosystemic models using several formalisms.
Keywords: system, complexity, bioinformatics, formalism, UML, object-oriented modelling
|Publications of the unit on Pasteur's references database|
|Office staff||Researchers||Scientific trainees||Other personnel|
|Roux-Rouquié, Magali, CNRS (DR2,email@example.com)||Grégory Sautejeau