Topic: New developments in inferring local and recent population dynamics from genomic data, and applications for the management of threatened populations and organisms of agronomic interest.
Doctoral school: University of Montpellier, GAIA
Dates: 1 October 2023 – 30 September 2026
CBGP supervisors: R. Leblois & A. Estoup

The development of agro-ecological approaches to the management of pests and beneficial organisms, as well as their vectors and natural enemies, requires a better understanding of the local demographic dynamics of their populations. Similarly, the management of threatened populations requires detailed knowledge of the demographic and genetic status of these populations: population size, fragmentation, dispersal, inbreeding, etc. Among the key factors to be characterised, population densities/sizes and dispersal characteristics at a small geographical scale, as well as their variations over the recent past, are often poorly understood but crucial for a better understanding of the dynamics of these populations. Spatialised genomic data contain information on these demographic parameters, but current analytical methods do not allow for the full utilisation of this information. Furthermore, the resulting estimates often relate to large spatial and temporal scales of limited practical interest. The aim of this project is to address this methodological shortcoming by developing new tools for estimating local demographic parameters, as well as their recent variations, from genomic data using spatialised demogenetic models.

Building on our current work on spatial population genetics models and the development of simulation-based inference methods, Ghislain will focus on:

1. finalise the implementation of temporal changes in demographic parameters within our spatialised genomic simulator, GSpace;

2. develop a set of new statistics or integrate artificial intelligence tools to summarise the genomic information relevant to the inference of parameters of interest;

3. combine these with high-performance simulation-based inference methods;

4. test the level of spatio-temporal complexity that can be accommodated depending on the type and quantity of available data.

In order to assess their practical value, these developments will be designed, tested and then applied in two quite different contexts:

1. the study and management of organisms of agronomic interest, through internal collaborations within the CBGP;
2. conservation biology, through the DevOCGen project and its collaborations.