Notation and ranking of subgraphs extracted from biological knowledge graphs

Type de poste
Niveau d'étude minimal
Durée du poste
Contrat renouvelable
Contrat renouvelable
Détails de renouvellement
Le contrat pourra être étendu de quelques mois à un an.
Date de prise de fonction
Date de fin de validité de l'annonce
Nom de la structure d'accueil

Campus scientifique des Aiguillettes
54500 Vandoeuvre les Nancy

Marie-Dominique Devignes
Email du/des contacts

The post-doctoral project concerns the analysis of complex interaction networks. The main ressource at our disposal is a huge heterogeneous graph database (provided by EdgeLeap for the FIGHT-HF program) that represents various types of interactions between various groups of elements : proteins, diseases, drugs, etc. One objective of the FIGHT-HF project is to exploit this ressource to identify new biomarkers characteristics of certain heart-failure mechanisms.
The queries on the main graph database most often return subgraphs such as the shortest paths between proteins or drugs of interest and a disease. In order to avoid manual inspection of all these subgraphs, some graph scoring should be defined in order to rank the subgraphs according to given points of view and to analyze first the most relevant ones. The graph scoring method should combine graph topological properties and any other properties attached as attributes to the graph nodes and edges, these latter properties being expressed in controlled vocabularies or ontologies.
Several graph scoring methods will be defined with the help of bioinformaticians, biostatisticiens et FIGHT-HF clinicians. The post-doctoral scientist will develop score calculation, and design and run evaluation studies, based for instance on already known biomarkers.

Equipe adhérente personne morale SFBI
Equipe Non adhérente