Machine learning and epigenotyping: a new lever for improving phenotype prediction in plants in an agroecological context

• Title: Machine learning and epigenotyping: a new lever for improving phenotype prediction in plants in an agroecological context.
• Doctoral student: Thomas WACQUET
• Affiliated unit: P2e University of Orleans, INRAE and Institut Agro Rennes-Angers
• Supervision:
  • Thesis supervisor: Stéphane MAURY - P2e University of Orleans, INRAE
  • Co-supervisor: Mathieu EMILY - Department of Statistics and Computer Science, Institut Agro Rennes-Angers, and IRMAR (Rennes Institute for Mathematical Research)
  • Thesis co-supervisor: Harold DURUFLE - BioFoRa, INRAE Orléans
• Doctoral school: Health, Biological Sciences and Chemistry of Living Organisms - SSBCV - No. 549, Orléans
• Project duration: 2025 - 2028

Project summary
The agroecological transitions of current agri-food systems are taking place in a context of increasing pressure due to climate change. To improve the sustainability of our agricultural systems in this context, it is essential to understand how individuals can adapt to environmental changes.
Genetic selection accounts for only part of phenotypic variation, and current models assume that the ranking of individuals is not affected by environmental conditions. However, it has been demonstrated in various species that environmental disturbances can influence the epigenome and phenotypic traits of organisms.
Faced with these challenges, which concern both cultivated plant species and farm animals, new phenotyping tools must be developed to monitor the suitability of the individual (plant or animal) and its environment (exposome). Epigenetic modifications, such as DNA methylation, are molecular markers that impact the phenotypic diversity of organisms.