Work Package 2

Tuesday 5 June 2012
by  daubin
popularity : 0%

Coordinator: Eric Tannier (INRIA, LBBE)

Contribution: All partners


Gene phylogenetic trees are usually constructed independently of each other. But some groups of genes do not actually evolve independently, either because they appear on a close range on a same chromosome, or because their participation in a common function constrains them to evolve together. A few existing models aim at uncovering correlations between gene histories: to detect co-evolving genes, Barker (2005) or Tuller (2010) use phylogenetic profiles, Loader et al (2007) use correlations between approximate relative divergence rates in a biogeographic context, while Ané et al (2007) propose a method to detect correlations between gene phylogenies that have exactly one gene per species. Gene phylogenies annotated with events of duplication, loss and transfer as obtained from WP1 provide an extremely rich data, whose analysis should reveal the amount and types of correlations between gene families.
The objective of WP2 is to gather, improve and build methodologies which aim at: (i) Detecting patterns of co-evolution among genes based on their annotated phylogenetic histories; (ii) Reconstructing the history of physical and functional relationships among genes; (iii) Using the co-evolution information to improve phylogenetic reconstruction.
WP2 can be seen as transversal to work-packages 1 and subsequent work-packages (3-6), because on one hand it will use the output of WP1, that is, a species phylogeny and reconciled gene trees, and provide a feedback to WP1 by adding crucial informations for the construction of reliable phylogenies, and on the other hand co-evolution and relations between groups of genes may be interpreted in terms of proximity on a chromosome, regulation, metabolic pathways, protein interactions or environmental constraints, providing a methodological background for WP3 to WP6.
This work package is intended at building a solid methodological foundation for many aspects of the project. We will deal with mathematical modeling and algorithmic problems. It will benefit from the long-lasting collaboration between algorithmicists in Montpellier (V. Ranwez and S. Bérard) and the LBBE, which hosts, in addition to strong methodological competencies, INRIA (National Computer Science Research Institute) researchers like the coordinator of this work-package.
The developments planned here are not necessary for most of the other work packages to start, because none of them is completely devoid of existing methodologies, but a common basis will facilitate their use in the context of WP1, as well as providing generic tools for the mathematical study of co-evolution.