• Title: "EAGLE - Artificial Intelligence for Translational Genetics in Agronomy," Université de Toulouse & INRAE;
  • PhD Student: Noémien Maillard;
  • Affiliated Unit: INRAE, GenPhySE (Génétique et Physiologie des systèmes d'Elevage);
  • Co-supervision:
    • Raphael Mourad, Université de Toulouse, MIAT (Mathématiques et Informatique Appliquées de Toulouse);
    • Julie Demars, INRAE, GenPhySE (Génétique et Physiologie des systèmes d'Elevage) ;
  • Doctoral School: SEVAB, Université de Toulouse.

Genomic and functional characterization of livestock animals appears to be a lever for agroecological transition through, among other things, the identification of genotype-phenotype relationships. Pan-genome association studies have identified thousands of variants associated with complex agronomic traits. However, the majority of these variants have been found in non-coding genomic regions, hindering the understanding of the underlying biological mechanism. Predicting molecular processes based on DNA sequence using deep learning methods represents a promising approach to understand the role of these non-coding variants. Traditional supervised learning requires DNA sequences associated with functional data for training, and the quantity of such data is highly limited by the finite size of the human genome. However, data augmentation approaches through orthology could significantly enrich the training datasets and thus improve the predictive capacity of models. The project aims to optimize and use deep learning models trained on human and murine data for porcine genomic data acquired in an experimental setup to identify the molecular architecture of phenotypes of interest. This strategy of translational predictive biology will contribute to improving agronomic traits for sustainable livestock farming.