[TYPES/announce] PhD position: Statistical learning to improve gene knockout predictions for metabolic networks based on abstract interpreation (BioComputing, Lille University, France)

[Cedric Lhoussaine]

The following position is again available for a very short period! If
you are interested, please answer as soon as possible at
bio-computing-apply at lists.gforge.inria.fr !

-- 
Cedric Lhoussaine
http://www.lifl.fr/~lhoussai

PhD: Statistical learning to improve gene knockout predictions for
metabolic networks based on abstract interpreation

Lille University, France (BioComputing team)

Prediction algorithms from computer science will become increasingly
relevant to guide the engineering of synthetic biological systems. An
important instance is the prediction of gene knockouts from metabolic
gene networks [2,6]. This problem is starting to gain industrial
relevance for the production of bio-active products by bacteria. An
important limitation of existing approaches, which are based on abstract
interpretation (a formal approach steming from program analysis) or
constraint based optimization, is the absence of quantitative
information on the strength of metabolic fluxes in such networks. We
propose to remedy the situation by inferring such quantitative
information from experimental data by applying statistical machine
learning techniques [5,7], and improving the previous prediction
algorithms based on abstract interpretation such that they can benefit
from the additional quantitative information.

The supervisory team: J. Niehren (INRIA) and C. Kuttler (Lille 1
University), both from the LIFL's BioComputing team. Local collaborators
will be, for statistical learning, M. Tommasi (Lille 3 University), and
for the biological aspects P. Jacques and F. Coutte (ProBioGem, Polytech
Lille). BioComputing and ProBioGem have already cooperated on related
topics for two years [1,3,4]. ProBioGem will recruit a PhD student in
biology, who will be cooperating with BioComputing's PhD student, on the
wet lab side.

    background : A Master's level with first-class academic credentials
is required, preferably in Computer Science, with knowledge on formal
methods. We might also consider candidates from biostatistics or maths.

    duration : 3 years, starting from Sept/Oct 2013.

    the project is fully funded, an open to students of any citizenship.

    mail contact bio-computing-apply at lists.gforge.inria.fr

    Please contact us AS SOON AS POSSIBLE !

Publications

[1]  F. Coutte, M. John, M. Bechet, M. Nebut, J. Niehren, V. Leclère,
and P. Jacques. Synthetic Engineering of Bacillus subtilis to
Overproduce Lipopeptide Biosurfactants. In 9th European Symposium on
Biochemical Engineering Science, Istanbul, Turkey, 2012.

[2]  A. Goelzer, F. B. Brikci, I. M. Verstraete, P. Noirot, P.
Bessieres, S. Aymerich, and V. Fromion. Re- construction and analysis of
the genetic and metabolic regulatory networks of the central metabolism
of Bacillus subtilis. BMC Systems Biology, 2(1):20+, 2008.

[3]  M. John, F. Coutte, M. Nebut, P. Jacques, and J. Niehren. Knockout
Prediction for Reaction Networks with Partial Kinetic Information:
Application to Surfactin Overproduction in Bacillus subtilis. In 3rd
International Symposium on Antimicrobial Peptides, Lille, France, June 2012.

[4]  M. John, M. Nebut, and J. Niehren. Knockout Prediction for Reaction
Networks with Partial Kinetic Information. In 14th International
Conference on Verification, Model Checking, and Abstract Inter-
pretation, volume 7737 of Lecture Notes in Computer Science, pages
355–374, Rome, Italy, Jan. 2013. Springer.

[5]  K. Murphy and S. Mian. Modelling gene expression data using dynamic
Bayesian networks. Technical report, UC Berkeley, 1999.

[6]  N. D. Price, J. L. Reed, and B. Ø. Palsson. Genome-scale models of
microbial cells: evaluating the consequences of constraints. Nature
reviews. Microbiology, 2(11):886–897, Nov. 2004.

[7]  K. Y. Yeung, K. M. Dombek, K. Lo, J. E. Mittler, J. Zhu, E. E.
Schadt, R. E. Bumgarner, and A. E. Raftery. Construction of regulatory
networks using expression time-series data of a genotyped population.
Proceedings of the National Academy of Sciences, 108(48):19436–19441, 2011.