Mathias Vandenbogaert


Business address :

Ecole normale Supérieure
46, rue d'Ulm
FR-75230 Paris,
France
Tel.: +33 (0)1 44 32 23 71
Email: Mathias.Vandenbogaert@ens.fr
WWW: http://algo.inria.fr/vandenbogaert

Personal Address:

104, rue Balard
75015 Paris,
France


Degrees:

End of scholarship's thesis in the laboratory of Biochemistry, Microbiology and Physiology:

Klonering en expressie van het Chlorobium limicola forma thiosulfatophylum cytochroom c-551 gen, deel uitmakend van een thiosulfaatverbruikend gencluster (Promotors: Prof. Dr. J. Van Beeumen, Dr. F. Verté).

Experience:

Computer related skills:

Teaching experience:

Publications:

Presentations:

Phd Study:

Subject Title: Extraction of functional signals in regulationally important genomic regions, and assessment of their statistical relevance. The PhD study is involved in the development of procedures aimed to search for signals located in upstream regions of coregulated genes. These regions are known to regulate the expression of these genes. In this context, the so-called structured motifs, as well as a word and his neighbors that share a common structure typical of a family are of particular interest. These signals (words), appear to be either overrepresented or avoided in specific genomes. Within this particular framework recent statistical and computational approaches are being handled with. This work will find applications in the determination of the secondary and tertiary structures of RNA molecules and proteines, as well in the determination of the relatedness of different species, according to the word usages in the genomic texts. This work has been the subject of various collaborations. M. Régnier and A. Denise (University of Orsay) are involved with the effective establishment of the formulas, for various structures of sets of words. A function library called QuickScore is currently under development, and will be containing the concepts of the approach. The studies on the relatedness of different bacterial species based on the current word-counting, statistical assesment and phylogenetics inference methods is a joint work with Dr. V. Makeev, and is supervised by Dr. M. Gelfand. This work, as well as the results obtained by Ms. M. Régnier and Mr. A. Denise in the statistical large deviations domain, which allow a very precise computation of probabilities on words, when exact calculations are very expensive or numerically unstable, have been applied in the search of functional signals in the zones of regulation (search for motifs applied to genomics).
Comments:
The effective calculation of the general explicit formulas for the assessment of the pertinence of genomic signals is simplified when the sets of words are structured. From a formal point of view, one is interested in particular in regular expressions and the approximate motifs which appear in many molecular biology problems. Within this particular framework, recent statistical calculations were applied to signals of regulation in the genomes of Bacillus subtilis, Arabidopsis thaliana and of Saccharomyces cerevisiae, as well as poly-adenylation signals in the human genome. The latter motifs were the subject of a statistical method that was presented making it possible to separate the artifacts words, e.g. neighbouring words with random composition, that are close to the required motifs. This approach makes it possible to obtain a concise list of high quality motifs, that is consistent enough to explain the over-representation of the motifs in the sequence. This work will also finds applications in the determination of secondary and tertiary structures of RNA and proteins.

Language Knowledge:



Sports: Volleyball, tennis



References :


Mireille Régnier

Serge Dulucq

Alain Denise

I.N.R.I.A. Rocquencourt

LaBRI - Université Bordeaux I

LRI, Équipe Bioinformatique.

Domaine de Voluceau

351 cours de la Libération

Université Paris-Sud

B.P. 105 78153 Le Chesnay Cedex

33405 Talence Cedex

91405 Orsay Cedex

France.

France.

France.

Mireille.Regnier@inria.fr

dulucq@labri.fr

Alain.Denise@lri.fr