Thierry Bastogne
ABOUT ME
Name: Thierry Bastogne
Sex: M
Age: 42 y
Status: Maried, 2 children
Native town: Vitry-le-François, Marne, FRANCE
Education: Habilitation à Diriger des Recherches
University: Nancy-Université
Research topic: Experimental design, modeling and control of biological systems (systems biology)
Position: Assistant professor
CONTACTS
Lab: CRAN (Centre de Recherche en Automatique de Nancy), CNRS, INRIA Lorraine
University : Nancy-Université
Address :
(1) Faculté des Sciences et Techniques, BP 70239, 54506 Vandoeuvre-lès-Nancy, France
(2) Centre Alexis Vautrin (Centre de Lutte contre le Cancer de Nancy), Avenue de Bourgogne F-54511 Vandœuvre-lès-Nancy, France
Email : thierry.bastogne@cran.uhp-nancy.fr
Tel : 03 83 68 44 73
Fax : 03 83 68 44 62
PHILOSOPHY
"My uniform Experience has convinced me that there is no other God than Truth"
(Mahatma Gandhi)
“La vérité ne se possède pas, ni ne se trouve, elle se cherche.”
(Condorcet, André Gide, Albert Jacquard)
LINKS
Links :
A few aspects of my research activity ...
We are interested by system identification problems involving errors in variables (input and output) but also timing errors. Potential applications lie in biology and biomedicine. Effects of the timing noise on the bias and variance of the output error are explicitly determined in [1] for a first-order linear model. A bounded-error parameter estimation approach is proposed as a suited solution to handle this problem. Application results are presented which emphasize the effectiveness of the methodology in such an experimental framework.
T. Bastogne, S. Mézières-Wantz, N. Ramdani, P. Vallois, L. Tirand, D. Bechet, and M. Barberi-Heyob. Parameter estimation of pharmacokinetics models in the presence of random timing errors. European Journal of Control, 14(2), 2008.
System identification with timing noise.
System identification of managed river reaches
Problem addressed in this study deals with the estimation of a time-delay from poorly informative data. This kind of problem particularly occurs in river reaches, managed for hydroelectricity production. Data were collected during a combined feedback/feedforward control carried out by human operators [2].
We proposed a Bayesian identification method, non-supervised and simple to implement, estimating jointly the time-delay and a finite impulse response (FIR). Experimental results show the effectiveness of the proposed method to estimate the river reach time-delay from data collected in imposed experimental conditions.
M. Thomassin, T. Bastogne, and A. Richard, “Identification of a managed river reach by a bayesian approach,” IEEE Contr. Syst. Technol., vol. 17, no. 2, pp. 353–365, 2009.
Object-modeling of interconnected dynamical systems
Modeling interconnected dynamical systems has been one the main topics of my research activities since 2005. I am particularly interested by the mathematical interpretation of the object-oriented modeling paradigm based on the Willems’ behavioral formalism of the systems theory. Two behavioral representations (complete and partial) of an object are proposed in [3]. Three object relationships, i.e. instantiation, composition and generalization are also examined in the behavioral framework. The implementation of the behavioral representations into the object-oriented language
Modelica is presented as well.
A unified description of networked dynamical systems, entitled multiport diagram, is proposed in [4]. The multiport diagram is an object-oriented model structure. It is shown that usual diagrammatic representations of networked systems, e.g. block diagrams, signal flow graphs, compartmental networks and bond graphs, are subclasses of the multiport diagram. By bridging the gap between these four network representations, this unified diagram should contribute to facilitate education of networked system modeling in systems and control theories.
Object-modeling and simulation of batch processes
One application of the multiport diagram (see previous paragraph) is the object-modeling of batch systems. The proposed modeling procedure is made up of five main steps: the hierarchical decomposition of the process into module classes, the modules modeling, their interconnection to form the multiport diagram, its implementation into a simulation platform and its calibration. The implementation of the multiport diagram is supported by the physical modeling language Modelica. Its simulation and calibration are performed into the environments Dymola and Matlab respectively.
Main effect design of experiments for photodynamic therapy
[3] T. Bastogne. Behavioral interpretation of the object-oriented paradigm for interconnected dynamic system modeling. International Journal of Systems Science, 38(4):319–326, April 2007.
[4] T. Bastogne. A unified representation for networked dynamical system modelling. Simulation Modeling Practices and Theory, 15(7):747–763, August 2007.
P. Mondrian
(1872-1944)
Multiport Diagram
T. Bastogne. A multiport object-oriented diagram for batch system modelling. Methodology and implementation. Simulation Practice and Theory, 12(6):425–449, October 2004.
Modeling responses of anti-cancer treatments
The cell membrane folate receptor (FR) is a molecular target for tumor-selective drug delivery, including delivery of photosensitizers for anticancer photodynamic therapy (PDT). Tumor selectivity of metatetra(hydroxyphenyl)chlorin (m-THPC), a photosensitizer used in PDT clinical trials, demonstrates a low tumor-to-normal epithelial uptake ratio. We report on the synthesis and on the photophysical properties of a m-THPC-like photosensitizer 1 conjugated to folic acid (compound 8). A main effect design based on a full factorial matrix was used to analyze additive and synergy effects on the selectivity response
J. Gravier, R. Schneider, C. Frochot, T. Bastogne, F. Schmitt, J. Didelon, F. Guillemin, and M. Barberi-Heyob. Improvement of m-THPC-like photosensitizer selectivity with folate-based targeted delivery. Synthesis and in vivo selective delivery study. Journal of Medicinal Chemistry, 2008.
Over the last few years, taking advantage of the linear kinetics of the tumor growth during the steady-state phase, tumor diameter-based rather than tumor volume-based models have been developed for the phenomenological modeling of tumor growth. In this study, we propose a new tumor diameter growth characterizing early, late and steady-state treatment effects. Model parameters consist of growth rhythms, growth delays and time constants and are meaningful for biologists. Biological experiments provide in vivo longitudinal data.
The latter are analyzed using a mixed effects model based on the new diameter growth function, to take into account inter-mouse variability and treatment factors. The relevance of the tumor growth mixed model is firstly assessed by analyzing the effects of three therapeutic strategies for cancer treatment (radiotherapy, concomitant radiochemotherapy and photodynamic therapy) administered on mice. Then, effects of the radiochemotherapy treatment duration are estimated within the mixed model. The results highlight the model suitability for analyzing therapeutic efficiency, comparing treatment responses and optimizing, when used in combination with optimal experiment design, anti-cancer treatment modalities.
T. Bastogne, A. Samson, R. Keinj, P. Vallois, S. Wantz-M´ezi`eres, S. Pinel, D. Bechet, and M. Barberi-Heyob, “Phenomenological modeling of tumor diameter growth based on a mixed effects model,” to be published in Journal of Theoretical Biology, 2009.
Response surface modeling for optimizing photodynamic therapy
Photodynamic Therapy (PDT) is based on the interaction of a photosensitizing agent (PS), light and oxygen, and a major difficulty is to select adequate treatment conditions. Few new PS are being developed up to the in vivo stage, partly because of the difficulty to find right treatment conditions.
Response surface methodology (RSM), an empirical modeling approach based on data resulting from a set of designed experiments, was suggested as a rational solution to select in vivo PDT conditions with a new peptide-conjugated PS targeting neuropilin-1. A Doehlert experimental design was selected to model effects and interactions of PS dose, fluence and fluence rate on the growth of U87 human malignant glioma cells xenografted in nude mice using a fixed drug-light interval. All experimental results were computed by the Nemrod-W® software and Matlab®. Intrinsic diameter growth rate, a tumor growth parameter independent to the initial volume of the tumor, was selected as response variable and was compared to tumor growth delay and relative tumor volumes. With only 13 experimental conditions tested, an optimal PDT condition was selected (PS dose: 2.80 mg/kg; fluence: 120 J/cm2; fluence rate: 85 mW/cm2). Treatment of glioma-bearing mice with the peptide-conjugated PS followed by the optimized PDT condition showed a statistically significant improvement in tumor growth compared with animals who received this PDT with the non-conjugated PS. RSM appears to be a useful experimental approach for rapid testing of different treatment conditions and determination of optimal values of PDT factors for any PS.
L. Tirand, T. Bastogne, D. Bechet, M. Linder, N. Thomas, C. Frochot, F. Guillemin, and M. Barberi-Heyob, “Response surface methodology: an extensive potential to optimize photodynamic therapy conditions in vivo,” International Journal of Radiation Oncology, Biology, Physics, vol. 75, no. 1, pp. 244–252, 2009.
System modeling of the photodynamic therapy
Since 2005, we have been working on the experimental modeling of the photodynamic therapy (PDT). This anticancer treatment is a multivariate and multiscale process. Its complete modeling requires to solve three main identification problems associated with the uptake, photoreaction and toxicity phases of the therapy. Main issues deal with experimental design, identifiability and parameter estimation. Moreover, we have shown that the light dosimetry problem in PDT could be expressed as a control issue which could bring new perspectives to the development of this therapy.
S. Dobre, T. Bastogne, M. Barberi-Heyob, D. Bechet, J. Didelon, and A. Richard, “System identification of the intracellular photoreaction process induced by photodynamic therapy,” in Proc. of the 16th IEEE Mediterranean Conference on Control and Automation. Invited session ’System Identification and Modeling of Biological and Biomedical Systems’ organized by T. Bastogne and B. Laroche, (Ajaccio, Corsica, France), June 25-27 2008.
S. Dobre, T. Bastogne, M. Barberi-Heyob, and A. Richard, “Practical identifiability of photophysical parameters in photodynamic therapy,” in Proc of the 29th Conf IEEE Eng Med Biol Soc., (Lyon, France), pp. 6633–6, August 2007.
T. Bastogne. Modélisation Expérimentale des Systèmes Dynamiques Interconnectés. Applications en Biologie Systémique. Habilitation à diriger des recherches, Université Henri Poincaré, Nancy 1, Juin 2008.
T. Bastogne, L. Tirand, S. Dobre, M. Barberi-Heyob, and A. Richard. Modélisation système de la thérapie photodynamique. Revue e-STA Sciences et Technologies de l’Automatique, 3(2), 2006.
T. Bastogne, H. Garnier, and P. Sibille, “A PMF-based subspace method for continuous-time model identification. Application to a multivariable winding process,” International Journal of Control, vol. 74, no. 2, pp. 118–132, 2001.
T. Bastogne, H. Noura, P. Sibille, and A. Richard, “Multivariable identification of a winding process by subspace methods for tension control,” Control Engineering Practice, vol. 6, pp. 1077–1088, september 1998.
T. Bastogne, Identification des systèmes multivariables par les méthodes des sous-espaces. Application à un système d’entranement de bande. Thèse de doctorat, Université Henri Poincaré, Nancy 1, novembre 1997.
Subspace identification of multivariate systems
Agent based model of in vitro Multicellular Tumor Spheroid Growth
In this study, the objective is not to develop a new agent-based model of tumor (see http://ms.izbi.uni-leipzig.de/) but to develop and assess statistical methods to analyse properties of agent-based models, e.g. identifiability of parameters, global sensitivity, etc. This simulator could also be used as an in silico benchmark in the identification of simplest probabilistic models or to test control strategies (anticancer treatments).
A. Orosanu, Agent based Simulations of the In Vitro Multicellular Tumor Spheroid Growth, Master of Science Thesis, Erasmus, Nancy-Université, University “Stefan cel Mare”, Suceava, June 2009. (MedicalDiseasev02_SD.pdf)
(Université de la Grande Région)
Die Systembiologie ist ein interdisziplinärer Ansatz, mit dem Ziel eines besseren Verständnisses und des Beherrschens der unterschiedlichen biologischen Funktionen auf allen Ebenen, von der Zelle bis zum ganzen Organismus. Fortschritte in Systembiologie basieren insbesondere auf Ermittlung, Sammlung und Kommentierung quantitativer Daten, die in mathematischen Modellen zur Erklären und Voraussagen wichtiger biologischer Funktion eingebunden werden. Die zukünftigen Herausforderung der Systembiologie finden sich hauptsächlich in der Medizin, der Pharmaindustrie und der Umweltforschung. Dieses Projekt vereint Forscher und Lehrende von vier Partneruniversitäten (Lüttich, Nancy, Luxemburg und Saarbrücken)
La biologie intégrative est une approche interdisciplinaire visant à mieux comprendre et maîtriser les différentes fonctions biologiques à tous les niveaux d’intégration auxquels elles se manifestent : depuis la cellule jusqu’à l’organisme tout entier. Les avancées de la biologie intégrative reposent, en particulier, sur la production, la collecte et l'annotation de données quantitatives qui pourront être utilisées dans des modèles mathématiques, explicatifs et prédictifs des grandes fonctions biologiques. Les perspectives d’enjeux de la biologie intégrative concernent essentiellement les secteurs médicaux, pharmaceutiques et écologiques. Le projet regroupe des enseignants-chercheurs de 4 universités partenaires (Liège, Nancy, Luxembourg et Sarrebruck)
