Thesis (M/F): MSI/IF image fusion by unsupervised learning

European Commission; France; Sur place; EUR 40 000 - 60 000; Plein temps

Résumé du poste

Une équipe de recherche à l'Institut de Pharmacologie et Biologie Structurale cherche un doctorant pour étudier l'imagerie spectrométrique afin de caractériser les microenvironnements tumoraux. Ce projet inclut la fusion de données d'imagerie à travers des techniques avancées d'apprentissage non supervisé, avec un soutien de deux groupes de recherche. Un profil en data science, traitement d'images, ou mathématiques appliquées est exigé.

Qualifications

• Motivation pour la recherche appliquée est essentielle.
• Expérience ou compétences en imagerie spectrométrique sont requises.
• Pas d'expérience biologique nécessaire, mais un intérêt pour le domaine est attendu.

Responsabilités

• Fusionner les données IF et MSI d'un même échantillon.
• Impliquer le pré-traitement, l'enregistrement et l'apprentissage de représentation pour réduire la complexité des données.
• Analyser les propriétés spatiales et spectrales.

Connaissances

Expérience algorithmique

Programmation

Expertise en imagerie spectrométrique

Formation

Master recherche ou diplôme d'ingénieur en data science, traitement d'images, mathématiques appliquées, ou informatique

Description du poste

Organisation/Company: CNRS
Department: Institut de Pharmacologie et Biologie Structurale
Research Field: Pharmacological sciences, Engineering
Researcher Profile: First Stage Researcher (R1)
Country: France
Application Deadline: 23 Jul 2025 - 23:59 (UTC)
Type of Contract: Temporary
Job Status: Full-time
Hours Per Week: 35
Offer Starting Date: 1 Oct 2025
Is the job funded through the EU Research Framework Programme? No
Is the Job related to staff position within a Research Infrastructure? No

Offer Description

The "Microenvironment, Cancer and Adipocytes" (MICA) team at IPBS uses mass spectrometry imaging (MSI) techniques to study small molecules and lipids, aiming to characterize tumor cell microenvironments and metabolism. MSI provides spectral data with high spectral fineness but limited spatial resolution, often combined with fluorescence microscopy (IF) for better spatial resolution. Previous work by Landry Blanc and collaborators has applied clustering methods to analyze this data. This thesis aims to explore unsupervised learning techniques for a more comprehensive fusion of MSI and IF modalities, as part of the ANR project SpaCELLM.

The doctoral student will be integrated into two research teams: The ADO group at ISAE-SUPAERO, focused on artificial intelligence, image processing, and vision; and the MICA team at IPBS, which researches the paracrine role of adipocytes in tumor progression and molecular mechanisms in normal weight and obesity conditions, studying cellular interactions via spatial analysis and molecular imaging.

Profile sought: Candidates should hold a research master's or engineering degree in data science, image processing, applied mathematics, or computer science. Strong academic background and motivation for applied research are essential. Experience or skills in mass spectrometry imaging are required. No biological experience is necessary, but interest in the field is expected.

The thesis will focus on merging IF and MSI data from the same sample, involving pre-processing, registration, and representation learning to reduce data complexity and analyze spatial and spectral properties. Approaches include spectral unmixing, low-rank approximation (e.g., NMF), and joint analysis of modalities, progressing through levels of complexity from constrained NMF informed by segmentation to unsupervised joint factorization methods, including deep learning techniques.

Required skills: algorithmic experience, programming, expertise in spectrometry imaging.