logo umr
Génétique Quantitative et Évolution - Le Moulon

Soutenance de thèse

 Arnaud DESBIEZ-PIAT

  -  14:00:00
 Visioconférence

The Dynamics of the Response to Selection under High Drift-High Selection: Insights from Saclay’s Divergent Selection Experiments for Flowering Time in Maize

Soutenance de thèse de doctorat de l’Université Paris-Saclay

École doctorale Sciences du végétal : du gène à l’écosystème (SEVE)

Vendredi 21 mars 2021 à 14h00

En visioconférence à l’URL :

https://eu.bbcollab.com/guest/823e5724db244a2c9197b9cfb884c3df

Veuillez vous identifier avec Prénom-NOM-Labo, SVP

Arnaud DESBIEZ-PIAT

Équipes BASE et GEvAD

Devant un jury composé de :

  • Guillaume Achaz, Professeur, Université de Paris, (MNHN, Paris) : Rapporteur
  • Luis-Miguel Chevin, Directeur de Recherche, CNRS (CEFE/CNRS, Montpellier) : Rapporteur
  • Marie-Anne Félix, Directrice de recherche, CNRS (IBENS, Paris) : Examinatrice
  • Laurence Moreau, Directrice de Recherche, INRAE, (GQE-Le Moulon, Gif-sur-Yvette) : Examinatrice
  • Kelly Swarts, Group Leader, Gregor Mendel Institute of Molecular Plant Biology (Vienne, Autriche) : Examinatrice
  • Christine Dillmann, Professeure, Université Paris-Saclay (GQE-Le Moulon, Gif-sur-Yvette) : Directrice de thèse
  • Maud Tenaillon, Directrice de recherche, CNRS (GQE-Le Moulon, Gif-sur-Yvette) : Co-encadrante, Invitée

Résumé

Understanding the adaptive dynamics sustaining phenotypic shifts and the limits to selection response are at the core of population and quantitative genetics. Based on the general assumption of random mating, large population size, weak selection intensity, theoretical models are well-suited for predicting observed responses in a broad range of parameters (with appropriate corrections). Yet, their use for understanding the long-term survival and adaptive dynamics of small selfing populations undergoing strong selection remains to be explored.

I focused on two experiments conducted for over 20 years under classic agronomic conditions: the Saclay Divergent Selection Experiments (DSEs) on maize flowering time, a highly complex trait with a wide mutational target. Each DSE consisted on applying High Selection (1% truncation selection) on two population of genotypes (early and late) derived from a single maize inbred line (He <0.5%) evolving under High Drift (Ne <4) and selfing (HDHS regime). These DSEs have been characterized by a strong response to selection. I asked several questions: How does the interplay between drift and selection influence the response to selection? What is the relative contribution of standing variation and de novo mutations to the observed phenotypic responses? What is the mutational dynamics under HDHS? How does phenotypic response at one trait impact other traits? What is the role of Genotype-by-Environment (GxE) interactions on phenotypic shifts?

I showed using a simulation that HDHS regime benefits from an enrichment of fixation of beneficial mutations revealing a limited cost of selection. From sequencing data, I detected hundreds to thousands of SNPs segregating between late and early populations, and tracked their fate through the DSE pedigrees. I evidenced two distinct adaptive phases: the former occurs through selection of beneficial standing mutations, whose fixation is likely delayed by clonal interference; the latter corresponds to the successive fixation of beneficial de novo mutations. I further analyzed field data collected for a subset of genotypes selected during the first 20 generations of the two DSEs, and evaluated for three traits (flowering time, leaf length and height) over two years. Interestingly, with a simple quantitative genetics model, I retrieved the two adaptive phases. Moreover, I found significant GxE interactions in all populations, that notably allowed for the decoupling of traits that are known to be correlated at the species level. These results provide new hints on how complex interaction between selection and drift triggers adaptation in small selfing populations, but also question the detection of adaptive variants in association mapping settings where the underlying pedigrees are fully controlled.