Postdoc position in computational Bayesian statistics

Here is an announcement I received that should interest potential postdocs (willing to come to Paris). The location is on the Orsay campus, south of Paris.

In the framework of the ANR-funded Metacoli project which aims at identifying the metabolic underpinnings of the lifestyle diversity in the E. coli species, Genoscope (the genomics institute of the French atomic energy commission) opens a 1-year post-doctoral position in Computational Bayesian Statistics. While Escherichia coli is mostly commensal bacterium of the gut, certain strains from this species are human pathogens causing multiple diseases, especially when escaping from the gut and colonizing other areas of the human body. The Metacoli project is specifically aimed at understanding how the operation of metabolism in certain pathogenic strains allows them to achieve this colonization, while other strains cannot. In the course of this project, a wealth of experimental data have been collected by two partners of the consortium (INSERM U722 “Ecology and evolution of microorganisms” and CNRS/INRA/Univ Paris XI mixt research unit “Plant Genetics”), while metabolic networks and models for 20+ strains have been reconstructed. Integrating these data using these models is an open challenge, to be tackled using Bayesian methods.

An adequate Bayesian model has been devised, partly based on past works published by Calvetti & Somersalo. Yet, no well-normalized analytical expression of the posterior is available, and one has to rely on sampling methods to gain insights into the posterior distribution of the metabolic concentrations and fluxes for each (strain, growth medium) combination. The post-doctoral fellow will be in charge of developing a suitable sampler to achieve this task, and will be involved in the downstream statistical analysis of the samples produced. Applicants should have experience with bayesian computing and strong programming capabilities. The successful applicant will work at Genoscope within the “Metabolic genomics” lab headed by Claudine Médigue, where a major part of the upstream modeling effort has already been conducted, and collaborate closely with the “Plant Genetics” unit, where part of the modeling expertise is located. The salary is approximately 2000 euros/month.

Prof.Christine Dillmann – Project coordinator Université Paris Sud & INRA & CNRS
dillmann [ à]
Dr. Claudine Médigue – Principal Investigator Genoscope – CEA/Life Sciences division
claudine.medigue [à]

4 Responses to “Postdoc position in computational Bayesian statistics”

  1. Actually I was not interested in that position, but in other positions like that in the near future and I want to improve my computational skills.

    Thanks anyway,


    • Sorry for the confusion Manoel! I cannot advise any book for advanced computing. Our book Monte Carlo Statistical Methods gives principles and methodology, rather than computing instructions, while the companion Introduction Introducing Monte Carlo Methods with R is only an introduction and anyway in R, which is not what’s needed for intensive computing. I would thus advise studying directly books about the computer languages like C, Python, as well as the computing environment (memory, processors, interfaces, data bases), including parallel processing (CUDA).

  2. Hello,

    What do you think it is “strong programming capabilites”?
    In order to enhance my prograaming capabilites and my experience with Bayesian Computing, do you suggest any book to start with? Your book, perhaps?

    I have a good knowledge of R, I’m learning C, and I do know hot to fit a multilevem model using WinBugs (I learned multilevel Bayes with Gelman’s book ARM).

    Thanks in advance,

    • Manoel, I am not connected with the call at all, so I have no idea of the level they are looking for. You should contact the persons in charge and ask for precisions. It is unlikely that starting training at this stage is going to work as they are desperate for a postdoc to apply today or, better, yesterday!!! My impression is that something like C would be what they want.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.