Search

On this website


Home > Directory > Boussau Bastien

Bastien Boussau
Chargé de recherche - CNRS
e-mail:
phone: 04 72 44 62 97
fax: +33 (0)4 72 43 13 88
UMR CNRS 5558 - LBBE
"Biométrie et Biologie Évolutive"
UCB Lyon 1  - Bât. Grégor Mendel
43 bd du 11 novembre 1918
69622 VILLEURBANNE cedex
Building: Mendel 2ème étage Office: 218

  • Scientific interests

    How have phenotypes been encoded in genomes throughout history?
    Addressing this question means tackling two problems: how have genomes evolved, and how are phenotypes encoded in genomes. The former requires novel models and algorithms; the latter requires analyzing extant genomes in a comparative framework to identify genomic characters that correlate with phenotypic characteristics. In the course of this research, I develop statistical approaches (Bayesian or maximum likelihood estimation) with high performance computing to study simultaneously large numbers of genomes.

    • +Studying evolutionary genomics with novel models and algorithms
  • For prospective students and postdocs

    Whether I advertise an open position below or not, I always welcome enquiries about possible research projects, from students and postdocs alike. These research projects can be aimed at a particular biological problem, or can be aimed at the development of a new probabilistic model or a new algorithm, possibly parallel. There are always interesting things to do, so please do not hesitate to send me an email with your research interest and current situation.

  • Collaborators

    I collaborate extensively with colleagues in the Bioinformatics and Evolutionary Genomics team: Thomas Bigot, Vincent Daubin, Laurent Duret, Manolo Gouy, Mathieu Groussin, Laurent Guéguen, Vincent Miele, Simon Penel, Eric Tannier. Topics include branch-heterogeneous models of sequence evolution and ancestral sequence reconstruction (Manolo Gouy, Mathieu Groussin), comparative genomics and database building (Laurent Duret, Simon Penel), integrative models of genome evolution (Thomas Bigot, Vincent Daubin, Vincent Miele, Eric Tannier). The latter work is done in very close collaboration with Gergely Szöllősi, ELTE-MTA Theoretical Biophysics Research group, Budapest, Hungary, and a former postdoc in our team.

    I also collaborate with Tracy Heath (UC Berkeley), Sebastian Höhna (Stockholm University), John Huelsenbeck (UC Berkeley), Michael Landis (UC Berkeley), Brian Moore (UC Davis) and Fredrik Ronquist (Stockholm University) on the development of the software RevBayes, a new computational framework for statistical phylogenetic analyses.

  • Grants and research funding

    Currently I am part of the Ancestrome project, which aims at reconstructing the genomes of extinct organisms, and use these to learn about ancient ecologies and ancient environments. Ancestrome is an "Investissement d’Avenir" project involving researchers from Lyon, Montpellier and Paris and headed by Vincent Daubin.

    A collaboration with John Huelsenbeck is funded thanks to the France-Berkeley funds, enabling us to visit each other’s lab in 2014.

    In the past, I have been funded by a Human Frontier Science Program Long Term Fellowship.

  • High Performance Computing

    Genomes contain up to several dozen thousand genes, and are accumulating at a fast pace in the databases. Their analysis with statistically sound methods therefore requires an important amount of computation.To achieve this, I aim to develop efficient algorithms, and I notably use parallel algorithms using the Message Passing Interface and OpenMP.

    These algorithms have been used to study genome evolution thanks to the machine Jade, at the GENCI center in France, where I had been awarded 3 milion hours. More recently, I obtained from the PRACE european research infrastructure 34 million hours to study genome evolution along the tree of life in collaboration with Vincent Daubin, Vincent Miele, Gergely Szöllősi and Eric Tannier. These hours will be used on the supercomputer Curie at the CEA.

  • Software

    In the course of my research, I develop software implementing probabilistic models for inference in a statistical framework.

    I mostly use C++ to implement efficient methods, and also rely on parallel algorithms, using the Message Passing Interface or OpenMP libraries. Such programs can then be used on High Performance Computing centers.

+Publications