Postdoctoral researcher (M/F) in representation learning for multi-primer environmental DNA data from soil communities

Job Description

Missions

Soils host a considerable share of terrestrial biodiversity and perform essential functions (decomposition, nutrient cycling, carbon storage). Yet this biodiversity remains largely unknown: most organisms are difficult to observe, their environmental drivers and interactions are poorly understood, and soils are rarely integrated into biodiversity models. Environmental DNA (eDNA) metabarcoding makes it possible to survey several compartments from a single sample, but raises the challenge of jointly analysing massive, high-dimensional and heterogeneous data. The project builds on the ORCHAMP observatory, which monitors soil communities along elevation gradients in the French Alps with six markers (eukaryotes, bacteria, fungi, insects, oligochaetes, springtails), coupled with climate, soil, land-use and vegetation data. The postdoctoral researcher will develop representation learning methods to link community composition to environmental conditions, first marker by marker, then by integrating several markers to study how soil compartments covary beyond their shared response to the environment.

Activity

The postdoctoral researcher will:
• review the literature on multimodal representation approaches and ordination methods in ecology
• take stock of the available eDNA and environmental data and set up reproducible analyses
• link eDNA and environment, marker by marker, comparing linear and non-linear approaches on held-out samples
• integrate several markers to study relationships between soil compartments and predict composition, trophic groups and diversity
• disseminate results through publications, presentations and open-source code
These directions are indicative: the postdoctoral researcher is encouraged to propose their own approaches.

Your Profil

Skills

• PhD in computational ecology, statistics, applied mathematics or artificial intelligence; knowledge of community ecology and macroecology is an asset
• experience in deep learning (required), knowledge of representation learning and ordination methods
• familiarity with eDNA and metabarcoding data strongly recommended
• Python, PyTorch or TensorFlow, scientific computing, handling of large and spatial datasets
• rigour in model validation and reproducibility
• autonomy, interest in interdisciplinary work, good level of scientific English

Application file: CV, cover letter. Applications must be submitted exclusively through the CNRS job portal. For scientific questions, please contact Wilfried Thuiller or Sara Si-Moussi.

Your Work Environment

The position is based at the Laboratoire d'Écologie Alpine (CNRS, Université Grenoble Alpes and Université Savoie Mont Blanc), one of the birthplaces of environmental DNA metabarcoding, within the BIOM team. The postdoctoral researcher will be supervised by Sara Si-Moussi and Wilfried Thuiller, in an environment combining ecology, bioinformatics and artificial intelligence, with access to the computing resources of the laboratory and the university.

Constraints and risks

Mainly computer-based work; occasional participation in ORCHAMP field campaigns is recommended.

Compensation and benefits

Compensation

from 3 041,58 depending on experience

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions