Postdoc position available at INRA Clermont-Ferrand -- LORIA/Inria Nancy
Grand Est 

Knowledge Discovery for biomarker identification 
(Knowledge Discovery based on Formal Concept Analysis, pattern mining
and preferences, for the identification of early predictive biomarkers
of diseases) 

Location: Clermont-Ferrand - Nancy 
Duration: 2 years 
Keywords: biomarker, prediction, Formal Concept Analysis, knowledge
discovery, multi-dimensional modeling. 

Description of the task. 

The goal of the project is to identify predictive (bio)markers of the
evolution of health status toward metabolic syndrome development (from
metabolomics signatures, socio-economic parameters and ``food habits''),
with the objective of building a model and determining whether the
integration of multidimensional parameters improves prediction. Finally,
this approach should allow to identify determinants of the evolution of
health status. In this project, the volume of data is very important and
data are as well heterogeneous (both numerical and symbolic). The
integration of large volumes of data can be guided by domain knowledge
and be supported by a data schema considered as a mediation system
(virtual integration needing correspondences between data sources). This
global schema can be based on a concept lattice and defined for
materializing the characteristics and the correspondences between data
sources. 
The concept lattice provides a classification structure that can be used
for various tasks, such as data indexing, information retrieval, data
mining, data modeling, and reasoning. The concept lattice is built
thanks to Formal Concept Analysis (FCA), which can be considered as a
symbolic method for knowledge discovery (KD). It is also planned to use
pattern mining methods for extracting frequent or rare patterns and
association rules as well. 

In this context, the post-doc fellow's research will consist in studying
the set of data to be analyzed from a theoretical and practical point of
view. The theoretical point of view consists in checking which symbolic
KD methods are appropriate for analyzing the data and which kind of
coupling with numerical KD methods could bring more useful results. 
The practical point of view consists in applying the given methods to
the data to be analyzed and to interpret the results. 
Algorithms for FCA, pattern mining and numerical KD methods will be
reused but new developments or adaptations are planned for carrying out
this project. 


Application: 
The candidate should prepare a detailed CV including a complete
bibliography, a motivation letter and recommendation letters as a single
pdf file. This file should be sent by email to both contacts below. 


Contacts: 

Estelle Pujos-Guillot, INRA (Institut National de la Recherche
Agronomique) 
UMR 1019 Human Nutrition Unit 
Research Centre of Clermont-Ferrand/Theix 
F-63122 St Genès Champanelle France 
Tel: +33 473 624 141 
Email: estelle.pujos@clermont.inra.fr 

Amedeo Napoli, LORIA (CNRS - Inria Nancy Grand Est - Université de
Lorraine) 
Équipe Orpailleur - Bâtiment B 
BP 239, F-54506 Vandoeuvre-les-Nancy 
Tel: +33 383 592 068 
Email: Amedeo.Napoli@loria.fr