Title	Postdoctoral Fellow
	Mail	This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
	Address	Symbiose
		INRIA/Irisa - Campus de Beaulieu
		35042 RENNES Cedex - France
	Tel	+33 2 99 84 75 78
	Fax	+33 2 99 84 71 71
Current Position	Postdoctoral fellow

   Last update : April 6th 2009

Member of the SITCON project: http://bioinfo-out.curie.fr/projects/sitcon/

SITCON is a systems biology project dedicated to the investigation of the Ewing tumors, pediatric tumors triggered by the formation of a chimeric fusion gene, so-called EWS-FLI1. This gene acts as an aberrant transcritption factor and deregulates the cells, leading to tumors with poor diagnosis.
 
Some of the overwhelmed functions deregulated by EWS-FLI1 are DNA duplication and reparation. This leads to chromosomal abnormalities such as duplicated or deleted genome regions.
 
Using CGH and expression data produced on 39 Ewing tumor biopsies, my work is dedicated to understand the links that relate gene copy disorders to gene expression disorders. We developped a methodology that confronts a model of key regulation pathways to experimental data using qualitative modeling.
 
So far, this methodology involves four main steps :
- Modeling a regulation network as qualitative equations,
- Encoding highthroughput data (expression and CGH microarrays) as qualitative variations,
- Resolving the qualitative equations,
- Extracting new knowledge from the consistency analysis and predictions.
 
This methodology combines modeling (step 1), data processing (step 2), deduction (step 3) and induction (step 4).  It aims at automating the logical reasoning of biologists when they confront experimental data to a model of regulatory pathways. In the long term, it can be used to investigate data warehouses encompassing knowledge data and experimental data in a systems biology fashion.

 
 

My work can be subdivided in three subprojects :
1. Designing a methodology to integrate CGH and expression data with a regulatory network
2. Conceiving new qualitative functions to model complex interactions (with Carito Guziolowski)
3. Discovering new insights on Ewing tumors

 

1. A methodology to integrate CGH and expression data with a regulatory network
During tumorigenesis, genetic aberrations arise and may deeply affect the tumoral cell physiology. It has been partially demonstrated that an increase of genes copy numbers induces higher expression; but this effect is less clear for small genetic modifications.
To study it, we propose a systems biology approach that enables the integration of CGH and expression data together with an influence graph derived from biological knowledge.

This work is based on 3 key ideas :
1) Interindividual variations in gene copy number and in expression allow to attack tumor varability and ultimately adresses the problem of individual-centered therapeutics.
2) Confronting post-genomic data to known regulations is a good way to check the soundness and limits of current knowledge.
3) The abstraction level of qualitative modeling allows integration of heterogeneous data sources.

We tested this approach on Ewing tumor data. It allowed the definition of new biological hypotheses that were assessed by random permutation of the initial data sets.

This work was the object of various oral communications and posters ([4,5,6,8]) and a publication to appear in the Proceedings of the BMIINT'09 meeting [3].

An extended version of this work is in preparation for publication in an international computational biology journal.
 

2. Defining new functions to model complex regulatory interactions (with Carito Guziolowski)
We have proposed a formal method to investigate the consistency of a regulatory model, represented by an influence graph, with large-scale differential data. The underlying idea of this approach is to automatize the biologist thinking about molecular interactions in order to identify lacks in the model that deserve specific corrections.

In previous works, a general interaction logical rule was used to address the question of consistency. Its efficiency was demonstrated on large-scale prokaryotic networks such as the E. coli transcriptional regulatory network. Regulatory networks on eukaryotes hold, however, more complex interactions since numerous post-translational interactions are involved in all critical functions.

We addressed the question of modelling steady states shifts induced by an experiment in complex networks involving post-translation interactions. The benchmark model is an interaction network describing major pathways implied in Ewing tumour development.

We have introduced specific ternary logical rules to model post-translational processes and we have incorporated these rules to an automatic reasoning framework based on dependence graphs. We used this framework to constrain the global behaviour of the network when confronted to qualitative effects obtained from transcriptome time series data.

We used these rules to reason over the influences that target cell cycle to identify which nodes could be manipulated, in order to provide paths that could revert the observed behaviour of the Ewing tumour phenotype.

This work was the object of an oral presentations [9]. A long paper was submitted to an international Systems Biology Conference [2].

3. Discovering new insight on Ewing tumors.
Using the methodology presented in paragraph 1., it is possible to explore the impact of inter-individual expression level and gene copy number variations via a gene regulation network.
We aim to adress several questions :
   - what is the impact of specific variations on the network ?
   - are there regular differential variations that may activate specific nodes or pathways in preferentially in metatstatic tumors w.r.t. non metastatic tumors ?
   - is the network robust to a wide range of possible variations ? If so, how to intereprate biologically the various validity ranges ?
   - for a given set of observations, is it possible to manipulate a limited number of nodes in order to revert specific pathways ?

We obtained results for these questions. We are currently working with our biologist partners in the Sitcon project for  theroretical and experimental validations.


 
Publications on the current project
 [1] S. Blachon, G. Stoll , C. Guzolowski , G. Pierron, S. Ballet , F. Tirode, O. Delattre, E. Barillot, A. Zinovyev, A. Siegel, O. Radulescu, A methodology to integrate CGH and expression data with a gene regulation network. In preparation.

[2] C. Guziolowski, S. Blachon , O. Radulescu, G. Stoll and A. Siegel, Designing logical rules to model complex interactions in biomolecular networks : An application to cancer model ling. Submitted to an international Systems Biology conference with peer review.

[3] S. Blachon, G. Stoll, A. Zinovyev, E. Barillot, O. Delattre, A. Siegel, O. Radulescu, Method for relating inter-patient gene copy numbers variations with gene expression via gene influence networks, Proceedings of the Workshop Biomedical Informatics and Intelligent Methods in the Support of Genomic Medicine (BMIINT’09), co-located with the 5th IFIP Conference on Artificial Intelligence Applications & Innovations (AIAI’09), Thessalonik, Greece, 23-25 April 2009, To appear.


Communications on the current project
 [4] S. Blachon, G. Stoll, A. Zinovyev, E. Barillot, O. Delattre, A. Siegel, O. Radulescu, Method for relating inter-patient gene copy numbers variations with gene expression via gene influence networks, Biomedical Informatics and Intelligent Methods in the Support of Genomic Medicine (BMIINT’09), Greece, 23-25 April 2009, Oral presentation.

[5] S. Blachon, C. Gudziolowski, G. Stoll, A. Zinovyev, O. Radulescu, M. LeBorgne, A. Siegel, Confronting regulation network models to CGH and microarray data : an application on Ewing tumors, IPG'08, Lyon, 19-21 November 2008, Poster

[6] S. Blachon, Anomalies chromosomiques et tumeurs d’Ewing, 6èmes Journées autour de la plateforme bioinformatique Genouest, Rennes, 21 October 2008, Oral Presentation

[7] G. Stoll, SITCON consortium, Presentation of SITCON project: Modeling signal transduction induced by a chimeric oncogene, ICSB'08, Göteborg, 22-28 August 2008, Poster

[8] S. Blachon, C. Guziolowsky, G. Stoll, T. Baumuratova, A. Zinovyev, M. Le Borgne, O. Radulescu, A. Siegel, Confronting a network to highthrouput data : a case study on Ewing tumors, Workshop on "Dynamical Modeling and Biological Network Simulation" colocated with JOBIM'08, Lille,3rd July 2008, Oral Presentation

[9] C. Guziolowski, S. Blachon, A. Siegel, Adding missing post-transcriptional regulations to a regulatory network, Workshop on "Dynamical Modeling and Biological Network Simulation" colocated with JOBIM'08, Lille,3rd July 2008, Oral Presentation

[10] S. Blachon, Bienvenue chez les CGHs, Séminaire Aubade, Vannes, 9-11 June 2008, AUBADE seminar of the Symbiose Project, Oral presentation

[11] T. Baumuratova, O. Radulescu, A. Siegel, C. Guziolowski, S. Blachon, G. Stoll, F. Tirode, K. Laud-Duval, O. Delattre, New approach to the construction of gene regulatory networks, SBMC'08, Poster
 

Previous work and publications

Before being involved in the SITCON project, my work was dedicated to SAGE data mining in the BM2A team leaded by Olivier Gandrillon at the CGMC and in the Turing team leaded by Jean François Boulicaut at the LIRIS.

My Phd Thesis is available here.

International Biology/Bioinformatics Journals

[1] J. Leyritz, S. Schicklin, S. Blachon , C. Keime , R.G. Pensa, C. Robardet , J. Besson , J-F. Boulicaut and , O. Gandrillon , SQUAT, a web tool to mine SAGE data. BMC Bioinformatics , 2008, 9:378.

[2] J. Klema, S. Blachon , A. Soulet, B. Cremilleux and O. Gandrillon, Constraint-Based Knowledge Discovery from SAGE Data. In Silico Biology. 8, 0014, (2008).

[3] S. Blachon, R.G. Pensa, J. Besson, C. Robardet, J-F. Boulicaut and O. Gandrillon, Clustering formal concepts to discover biological ly relevant knowledge from gene expression data, In Silico Biology, (2007). 

[4] C. Becquet , S. Blachon, B. Jeudy, J-F. Boulicaut and O. Gandrillon, Strong association rule mining for large gene expression data analysis : a case study on human SAGE data , Genome Biology 3(12) :research0067.1-0067.16, (2002).

Book Chapter

[5] S. Blachon, C. Robardet , J-F. Boulicaut et , O. Gandrillon , Extraction de connaissances dans les donnees d’expression SAGE humaines, in Informatique pour l’Analyse du Transcriptome, Hermès Science, Traite IC2, (2004).

Proceedings of international Computer Science Conferences with lecture comittee

[6] J. Klema, A. Soulet, B. Cremilleux, S. Blachon and O. Gandrillon, Mining Plausible Patterns from Genomic Data, 19th IEEE International Symposium on Computer-Based Medical Systems (CBMS’06), pp. 183-188, IEEE Computer Society Press, Salt Lake City, Utah, June. (2006).

[7] C. Hébert, S. Blachon, B. Cremilleux, Mining delta-strong characterization rules in large SAGE data, In ECML/PKDD’05 Discovery Challenge on gene expression data colocated with the 9th European Conference on Principles and Practice of Knowledge Discovery in Databases PKDD’05 , pages 90-101, Porto, Portugal. (2005). 

[8] F. Rioult, C. Robardet, S. Blachon, B. Cremilleux , O. Gandrillon, J-F. Boulicaut, Mining concepts from large SAGE gene expression matrices, Proceedings of the 2nd International Workshop on Knowledge Discovery in Inductive Databases KDID’03 co-located with ECML-PKDD 2003, Catvat-Dubrovnik (Croatia), September 22. (2003).