Anne Jeannin Former Projects
Immune system modelling
(State University of New York, Stony Brook) During my postdoctoral training, I have been interested in modeling the immune system. In particular, I focused on the reactions in germinal centers, emerging structures in secondary lymp nodes during the immune response. In those tissues, lymphocytes B receptors improve their affinity with foreign antigen (the secreted form of those receptors are the antibodies, or immunoglobulins). The main idea was to consider this process of affinity maturation across multiple scales: at the mesoscopic scale (spatial aspects of germinal centers: migratory patterns of B lymphocytes; cell / cell interactions between T and B lymphocytes and follicular dendritic cells). At the sub-cellular sclale, a mutational model is applied on the B cells' receptors in order to take into account the mutations of those sequences under the action of the AID enzyme (Activation-induced cytidine Deaminase) and of the error-prone DNA repair pathways. The affinity of the immunoglobulin are re-evaluated after mutation (through protein-protein interactions) and a selection process allow to "export" (or not) B cells outside the germinal center. This kind of approach allows to ask questions about the diversity of the gene repertoire that constitutes the immunoglobulin, the clonal diversity of germinal centers, or the distribution of the motifs on the immunoglobulin sequences that are preferentially targeted by AID. At the mesoscopic scale, we can ask questions about the non-heterogeneity of endosomes in B cells during mitosis, or how a meta-population approach can influence the immune response.
I also contributed to the finalization of a software dedicated to NGS immunoglobulin sequences alignment.
Numerical simulation of biological tissue morphogenesis
(University of Brest, France) While working on my PhD degree, I proposed models and tools to simulate tissue morphogenesis in silico. Technologies used in molecular and cellular and biology generates a tremendous amount of data. The conception of integrative and predictive models can allow researchers to better understand biological systems. Multi-cellular systems exhibit mainly two levels of complexity: the first one concerns the potentially large amount of cells they contain, which requires a considerable computing power when this point is addressed through individual-based approach; the second level concerns the diversity of biological cells' behaviors. This level requires specific algorithms, for example to deal with complex behavior such as mitosis. The conception of 1) models that integrate biological data and 2) dedicated algorithms adapted to heterogenous and multi-core devices make it possible to solve, at least to some extent, these two levels of complexity. In order to numerically study both healthy and pathological tissue development, we propose two elements. The first is a biomechanical cell model that includes the behaviors involved in tissue morphogenesis (mitosis, differentiation, adhesion, migration, cell-cell signalling and apoptosis). The second element is a parallel simulator that relies on a non-specialised software architecture and on dedicated data structures and algorithms used to benefit from the power of multi-core hardware. In this document, we present several case studies that gives some validation elements of both our model and our simulator.
Interaction dynamics of agents in virtual environment
(University of Brest, France) The research work I did for my master's thesis concerned the autonomy and interactivity of agents evolving in virtual environment. Existing approaches to model agent's behavior are based on rules or on reasoning, but in that case, agents may lack reactivity when facing perturbation of the environment. A way for agents to be more reactive is to generate rule dynamically, but we tried yet an alternate approach in which it was not necessary to explicit agent's behavior, even dynamically, while allowing them to be adaptative and reactive. We used principles from ecological cognitive psychology, in particular a principle stating that coupling dynamical systems must allow the emergence of beheviors thanks to an adaptative and reactive system. We applied this principle with a basic task: ball bouncing. We chose this task because it has been well characterized in cognitive psychology. Despite the simplicity of the task, the original approach we developed allowed us to observe emerging reactive behaviors without any need to explicit behavioral rules, even dynamically.
Publications
- Peer-reviewed international journals
[1] Jeannin-Girardon A., Ballet P. and Rodin V., « Large scale tissue morphogenesis simulation on he- terogenous systems based on a flexible biomechanical cell model », IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB), 12(5) :1021-1033, Sept.-Oct. 2015.
[2] Jeannin-Girardon A., Ballet P. and Rodin V., « A software architecture for multi-cellular system simulations on graphics processing units », Acta Biotheoretica, vol. 61, no. 3, pp. 317–327, 2013a.
[3] Buche C., Jeannin-Girardon A. and De Loor P., « Simulation theory and anticipation as a basis for interactive virtual character in an uncertain world. Application to a human-virtual characters interaction for juggling », Computer Animation and Virtual Worlds (CAVW), Computer Animation and Social Agents (CASA’11) Special Issue, 22(2-3) :133-139, 2011.
- Invited international conferences
[4] Rodin V., Jeannin-Girardon A., Sarr A., Rivière J., Fronville A. and Ballet P., « A multi-agent approach for virtual tissue morphogenesis », 2nd Symposium on Complex Biodynamics & Networks, Tsuruoka (Japan), 11-13 may, 2015.
- Peer-reviewed international conference proceedings
[5] Jeannin-Girardon A., Ballet P. and Rodin V., « In silico study of mechanical stresses at the cellular level during tissue development », 13th IEEE International Conference on BioInformatics and BioEngineering (BIBE 2013), 2013c.
[6] Jeannin-Girardon A., Ballet P. and Rodin V., « An effcient biomechanical cell model to simulate large multi-cellular tissue morphogenesis : application to cell sorting simulation on GPU », 2nd International Conference on Theory and Practice of Natural Computing (TPNC 2013), LNCS volume 8273, pages 96-107, 2013b.
- Book chapter, thesis
[7] Jeannin-Girardon A., Développement d’un modèle logiciel de cellule sur processeurs multi-cœurs pour la simulation de morphogenèse de tissus, thèse de doctorat, Université de Bretagne Occidentale, 2014.
[8] Ballet P., Pothet A., Misevic G., Jeannin-Girardon A., Fronville A. and Rodin V., Une approche multi- agent pour la simulation en biologie cellulaire, Dans Le vivant discret et continu. Modes de représentation en biologie théorique, Eds. Matériologiques chap. 6, p. 155–194, 2013.
[9] Jeannin-Girardon A., Couplage de systèmes dynamiques pour l’émergence de comportement en environnement virtuel : application au rebond de balle, mémoire de master recherche en informatique, Université de Bretagne Occidentale, http://dumas.ccsd.cnrs.fr/docs/00/63/64/31/PDF/Jeannin-Girardon.pdf, 2011.
- National conference proceedings / other communications
[10] Jeannin-Girardon A. and Rodin. V, « Gestion efficace des ressources mémoire et de calcul pour l'exécution de systèmes multi-agents sur architectures parallèles avec OpenCL », Compas'2016, Conférence d'Informatique en Parallélisme, Architecture et Système, Session parallélisme #5: Support d'exécution, P5.1, 8 pages, Lorient (France), 5-8 juillet 2016.
[11] Jeannin-Girardon A., Ballet P. and Rodin V., « Simulation biomécanique de cellules virtuelles. Étude in silico de contraintes mécaniques pendant le développement de tissus », Dans 2ème séminaire de l’équipe IHSEV, juillet 2013.
[12] Guevel E., Jeannin-Girardon A. and Dezan C., « Du paramétrage de la granularité du calcul et de la localité des données des implémentations sur GPU - Expérimentations OpenCL », Dans Colloque annuel du GDR SOC-SIP, France, juin 2013.
[13] Jeannin-Girardon A., Ballet P. and Rodin V., « Structure de données optimisée pour la conception de simulation de systèmes multi-cellulaires sur architecture GPU », 2ème journée des doctorants de l’ED SICMA, sept. 2012.
[14] Jeannin-Girardon A., Ballet P. and Rodin V., « Architecture logicielle pour la conception de simulation de systèmes multi-cellulaires sur GPU », 32ème séminaire de la Société Francophone de Biologie Théorique (SFBT 2012), p. 25–26, 2012. [Prix Pierre Delattre 2012 (best presentation) awarded jointly to 3 PhD candidates by the Francophone Society of Theoretical Biology]