Mots-Clés

Modélisation du métabolisme machine learning Cameline

Description

Research unit

Currently counting over 150 staff members including non-permanent personnel, UMR 1332 BFP is a partnership between INRAE (Divisions of Biology and Plant Breeding and Plant Health and Environment) and the University of Bordeaux. It constitutes a major pillar of plant biology research in Nouvelle- Aquitaine. In parallel to its research activities, BFP has a strong involvement in teaching and training in the Plant Sciences field but also in Microbiology and Biochemistry. It has strong partnerships and a high international visibility, as indicated by the coordination of several EU-funded projects or collaborative networks, the establishment of an International Associated Laboratory with the University of Tsukuba in Japan (LIA FReQUenCE) and participation in a number of international projects. BFP also shows strong partnerships in several areas with the industry or with grower organisations.

The Meta team is a multidisciplinary team (analytical chemistry, biochemistry, molecular biology, physiology, statistics and bioinformatics) using Systems Biology approaches to understand metabolism and the way it participates in the performance of plants, whether in terms of yield or adaptation to abiotic and biotic stress. The team generates large datasets to develop, parameterise and validate predictive models. Bottom-up mathematical models (from mechanisms to the phenotype) are used to understand how the fluxes and concentrations of metabolites are controlled. Top-down models (from the phenotype to mechanisms) are used to look for metabolic markers associated with performance, the first step towards identifying the underlying mechanisms. While in recent years the team has focused on central metabolism, now it is also interested in redox metabolism as well as secondary metabolism, in particular because of their involvement in adaptation to the environment. The Meta team also hosts the Bordeaux Metabolome Platform that enables a wide range of targeted and untargeted metabolomics.

Context:

Climatic variability and extreme weather events are increasingly impacting crop yield and value. The European H2020 Project UNTWIST addresses this challenge and will take advantage of the climate- resilient, traditional European oilseed crop camelina to provide a mechanistic understanding of successful adaptation strategies to drought and heat stress. Moreover, it will develop predictive models and robust markers for crop performance for variable environments. UNTWIST ́s long term aim is to improve crop resilience and yield stability in changing, challenging climates; its legacy will contribute towards increasing the sustainability of European agriculture.

The Meta Team is leading WP4, which aims to study the links between metabolism and plant performance, by integrating a wide range of datasets ranging from omics data to whole-plant traits.

Objectives and work program of the post-doctoral fellow

The post-doctoral fellow will use both bottom-up and top-down modelling approaches. He/She will perform genome-scale metabolic network reconstruction of 5 different camelina lines based on deep sequencing of the different lines as part of the project. Once the reconstruction is performed, he/she will be involved in the analysis of the networks, including the integration of different omics dataset, with the aim of better understanding the phenotypic differences between camelina lines.

Besides, the post-doctoral fellow will have the opportunity to use and/or propose machine learning approaches to perform top-down modelling and predict phenotypic traits from the different omics data available, using both supervised and unsupervised approaches.

The post-doctoral fellow will be involved in a modeling group, within the Meta team, made up of several researchers, another post-doctoral fellow over the same three years period, and phD students.

REQUIRED QUALIFICATIONS

Candidate skills: Experience in genome-scale metabolic networks reconstruction and/or modelling is essential. Knowledge, even basic, in machine learning would be a valuable asset. Interest in plant metabolism would also be appreciated. The candidate must be able to work in a network and to report on his/her results (oral and written communications).