Subject: Genomic prediction of adaptation: statistical developments and application to the crop pest Drosophila suzukii.
Dates: 1 October 2024 – 30 September 2027
CBGP supervisors: S. Boitard & M. Gautier
University : Institut Agro, Montpellier

By combining genomic and environmental data collected from a wide range of populations presumed to be locally adapted, we can identify which alleles influence adaptation to a given environment and thus predict the adaptive potential of a genotyped target population in a new environment. An important application of this approach is to predict the risk posed by a population of a (pest) species of interest in a given environment, in particular the risk of an invasive population establishing itself in a new geographical area based on its current (or future) climate, the potential damage caused to a given host plant, or its ability to resist control methods.

In this thesis, we will explore two alternative approaches for making such predictions: the hierarchical Bayesian model implemented in the Baypass method (Gautier 2015), which will need to be extended for predictive purposes, and deep neural networks. We will apply these approaches to the genomic prediction of adaptation to different fruits in the crop pest Drosophila suzukii, drawing on data generated during various recent projects.