CJS INRA ELIGIBLE *

Date : 4/3/2010

Modelling the topology of a host-parasite interaction network with statistical models for heterogeneous random graphs.

Laboratory 1
Mathématique et Informatique Appliquées
UMR 518 AgroParistech/INRA
16 rue Claude Bernard, 75005 Paris
Unit/Lab Director : Stéphane Robin
http://www.agroparistech.fr/mia/doku.php

Laboratory 2
BIOGECO, BIOdiversité, Gènes et Communautés
UMR1202 INRA/Université Bordeaux1
69 route d'Arcachon, 33612 Cestas Cedex
Unit/Lab Director : Antoine Kremer
http://www.pierroton.inra.fr/biogeco/presentation-en.html

PhD Supervisor 1
Jean-Jacques Daudin
email : Cet e-mail est protégé contre les robots collecteurs de mails, votre navigateur doit accepter le Javascript pour le voir
phone : +33 1 44 08 16 67

PhD Supervisor 2
Corinne Vacher
email : Cet e-mail est protégé contre les robots collecteurs de mails, votre navigateur doit accepter le Javascript pour le voir
phone : +33 5 57 12 27 24

Subjects / Tools-Methodologies
1 Identifying Modules in Graphs / Clustering and Mixture Models
2 Topology of an Ecological Network / Statistical Analysis of Complex Data Sets
3 Impact of Global Change / Metagenomic Data Analysis

* CJS INRA eligible : available funding
This project may be selected to be founded by a 5-year grant from INRA (Early stage scientist contract)

Summary of labs' interests

Lab1: The UMR 518 is composed of statisticians and computer scientists working on learning methods and statistical models for Biology. The objective is to propose new models and new algorithms for the analysis of genomics or environmental data. The members of the UMR are particularly involved in complex models with latent variables, model selection and variationnal and bayesian inference. (see )

Lab2: The research activities conducted by BIOGECO are aimed at elucidating the mechanisms shaping biological diversity in terrestrial ecosystems at different hierarchical levels: communities, species, populations, genes. The main objective of BIOGECO is to provide an integrated analysis of diversity by adopting a threefold strategy: (1) integrated analysis of diversity from genes to communities, (2) geographical and spatial analysis of diversity from stands to landscapes and (3) distribution of species and historical reconstruction of diversity from the holocene to today (see http://www.pierroton.inra.fr/biogeco/presentation-en.html)

Summary of project

The main objective of the thesis is to elucidate the topology of a host-parasite (tree - fungus) ecological network by using the last developments of the statistical models for heterogeneous random graphs. The impact of global change on the tree - fungus network will also be studied through the statistical analysis of the data arising from a metagenomic experiment.
More details in http://www.agroparistech.fr/mia/doku.php?id=accueil:stages#post-doctoral_position.
References for the statistical part
• Daudin, J-J., Pierre, L., Vacher, C. Model for heterogeneous random networks using continuous random variables and an application to a tree-fungus network (2010). Biometrics,early view.
• F. Picard, V. Miele, L. Cottret, J.J. Daudin and S. Robin, Deciphering the connectivity structure of biological networks using MixNet. BMC Bioinformatics 2009, 10(Suppl 6).
• J.J. Daudin, F. Picard, S. Robin (2007) A mixture model for random graphs, Statis. Comput.18(2), 173-183
• Mariadassou M. and Robin S. (2007) Uncovering latent structure in valued graphs: a variational approach. Technical Report 10, SSB.
References for the ecological part
• Vacher, C., Daudin, J-J, Piou, D., Desprez-Loustau, M-L. 2010. Ecological integration of alien species into a tree-fungus network. Biological Invasions, in press
• Vacher C, Piou D, Desprez-Loustau M-L (2008) Architecture of an antagonistic tree/fungus network: the asymmetric influence of past evolutionary history. PLoS ONE 3:e1740