En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR7503-MALSMA-001 - Post-doctorant/post-doctorante en apprentissage automatique et modélisation structurale appliqués à l'immunogénécité des molécules HLA (H/F)

Post-doctoral position in machine learning and structural modeling applied to immunology (M/F)

This offer is available in the following languages:
- Français-- Anglais

Date Limite Candidature : jeudi 23 octobre 2025 23:59:00 heure de Paris

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler

Informations générales

Intitulé de l'offre : Post-doctoral position in machine learning and structural modeling applied to immunology (M/F) (H/F)
Référence : UMR7503-MALSMA-001
Nombre de Postes : 1
Lieu de travail : VANDOEUVRE LES NANCY
Date de publication : jeudi 2 octobre 2025
Type de contrat : Chercheur en contrat CDD
Durée du contrat : 18 mois
Date d'embauche prévue : 23 octobre 2025
Quotité de travail : Complet
Rémunération : gross salary : Starting from €3081, depending on professional experience.
Niveau d'études souhaité : Doctorat
Expérience souhaitée : Indifférent
Section(s) CN : 55 - Sciences et données

Missions

This contract concerns part of an ANR project (EPIHLA2), whose objective is to extend the HLA-Epicheck model, originally developed within the framework of a PhD thesis, and to implement new deep learning approaches to assess donor–recipient compatibility in organ transplantation. HLA-Epicheck is a predictive model of the antigenicity of polymorphic amino acids on the surface of HLA antigens, relying on dynamic structural data.

Activités

Four tasks are identified.
Task 1: Extension of the coverage and performance of the HLA-Epicheck model through the addition of new antigens. This also includes optimizing the values of certain model parameters.

Task 2: Evaluation of donor/recipient (D/R) epitope compatibility using two methods. HLA-Epicheck does not allow quantification of structural differences between two epitopes. The first option is to use the descriptors exploited by HLA-Epicheck to measure structural similarity. The second option relies on learning a new latent structural representation of the surface of an HLA antigen by combining a variational autoencoder (VAE). For both options, epitope-level similarities must then be aggregated at the antigen level and ultimately at the D/R pair level to estimate D/R compatibility (objective of Task 4)

Task 3: Consideration of hidden polymorphic amino acids (AAs).
To account for all potential epitopes of a donor (or recipient), it is important not to exclude non-polymorphic surface AAs, as these may correspond to epitopes induced by polymorphisms that are not directly exposed on the surface.

Task 4: Prediction of D/R compatibility based on the complete epitope load.
In order to assess compatibility between a donor and a recipient, it is essential to consider all pairs of class I and class II antigens. A model must then be trained using as input the donor and recipient epitope loads in order to predict their compatibility. Data from a cohort held by the project coordinator will be used to build this model.

Compétences

Python programming.
Analysis and development of ML and DL models using standard data science libraries. Data engineering from molecular dynamics simulations and public databases.

Contexte de travail

Loria laboratory, Capsid group in the fifth department (D5)

Contraintes et risques

Admission to the LORIA laboratory, hosted by Inria, is subject to approval by the FSD.