Thèse Génomique du Paysage de l'Adaptation à un Agent de Biocontrôle chez un Insecte Ravageur du Pommier H/F

Doctorat.Gouv.Fr

Établissement : Avignon Université École doctorale : Agrosciences et Sciences Laboratoire de recherche : Plantes et Systemes de cultures horticoles Direction de la thèse : Bertrand GAUFFRE ORCID 0000000154318098 Début de la thèse : 2026-10-01 Date limite de candidature : 2026-05-18T23:59:59 CONTEXTE
Les adaptations des ravageurs aux méthodes de lutte constituent un exemple emblématique d'évolution en action. Longtemps associées aux pesticides, les résistances touchent également les agents de biocontrôle, remettant en question leur durabilité. Ces adaptations ne peuvent être comprises comme de simples réponses locales à la sélection : leur dynamique est fondamentalement spatiale. À l'échelle des paysages agricoles, cette dynamique spatio-temporelle résulte de l'interaction étroite entre pression de sélection (liée aux pratiques culturales) et la dispersion des individus, qui redistribue en permanence les variants génétiques entre populations et est fortement contrainte par la structure des paysages, déterminant leur connectivité. Dans ce contexte, comprendre et prédire les dynamiques spatio-temporelles des résistances nécessite une approche intégrative reliant génomique, démo-génétique, analyses spatiale et télédétection, afin d'identifier les échelles pertinentes de surveillance et de gestion, condition essentielle pour assurer une efficacité durable des produits de biocontrôle.
Dans ce cadre, ce projet se concentre sur l'adaptation d'un ravageur majeur du pommier, le carpocapse des pommes (Cydia pomonella), à un agent de biocontrôle, le virus de la granulose (CpGV). Utilisé depuis les années 1990 pour réduire la dépendance aux pesticides et proposer une alternative sans impact sur l'environnement, le CpGV constitue un pilier de la protection des vergers. Toutefois, depuis leur première détection en 2005, l'adaptation aux produits à base de CpGV est en augmentation, compromettant l'efficacité de cette méthode de lutte largement utilisée en agriculture biologique et conventionnelle.
La thèse s'appuie sur des résultats récents ayant permis d'identifier les bases génétiques et des marqueurs moléculaires associés aux adaptations de C. pomonella au CpGV. Elle repose également sur le génotypage de plus de 6000 marqueurs génétiques (neutres et associés aux résistances) réalisé sur des pools d'individus issus de plus de 100 vergers à l'échelle d'un bassin de production du sud-est de la France (site atelier Basse Vallée de la Durance : https://site-atelier-basse-vallee-durance.fr/) et plus de 300 vergers à l'échelle nationale. Ces données seront complétées au cours de la première année de thèse. Par ailleurs, des données à l'échelle du paysage couvrant la distribution des vergers, l'occupation du sol et les pratiques agricoles sont déjà disponibles à l'échelle du site atelier BVD. L'ensemble de ces données sera mobilisé par le ou la futur·e doctorant·e.

OBJECTIFS ET QUESTIONS SCIENTIFIQUES
L'objectif central de ce projet de thèse est d'étudier comment les pressions de sélection, locales et à l'échelle du paysage, interagissent avec la dispersion et la structure des paysages agricoles pour déterminer les dynamiques spatio-temporelles de la résistance au CpGV chez C. pomonella, à plusieurs échelles spatiales. Cette problématique sera abordée à l'aide d'approches de génomique du paysage et de démo-génétique.
Les principales questions de recherche incluent : Quelle est la structuration génétique des populations de C. pomonella à différentes échelles spatiales, et que révèle-t-elle sur les flux de gènes et la connectivité ? Quels facteurs paysagers, environnementaux et agronomiques (densité et distribution des vergers, climat, intensité des traitements) influencent la fréquence et la distribution des résistances ? Est-ce que la prise en compte de la connectivité permet de mieux expliquer et de prédire la distribution des résistances ? Dans quelle mesure les séries d'échantillons historiques permettent-elles de reconstruire la dynamique spatio-temporelle des résistances ? À quelles échelles spatiales et temporelles organiser la surveillance et la gestion collective afin d'optimiser la durabilité des stratégies de biocontrôle basées sur le CpGV ? Understanding the processes that drive the emergence and spread of resistance in agricultural landscapes is critical to preserving the efficacy and durability of control methods. This need is particularly timely in the context of agroecological transition as the reduction of authorized active substances increases reliance on alternative solutions, including biocontrol methods, which are also affected by pest adaptation. Identifying whether populations are resistant or susceptible is a first step, but it is insufficient to predict the increase and spread of resistance within and among production regions, or to identify the levers to contain it. Spatio-temporal dynamics of resistance depend not only on the variability of local selection pressures but also on demographic processes governing pest populations (Lenormand, 2002; Miller, 2017). Population size and dispersal depend on the components of the environmental matrix, through biotic features (e.g., host plants, landscape structure) and abiotic drivers (e.g., farming practices, climate). In agricultural landscapes, this matrix can be highly dynamic in space and time due to the diversity, abundance, distribution, and phenology of host plants, as well as the diversity of farming practices. It is therefore essential to quantify functional connectivity by clarifying how the environmental matrix modulates dispersal and the resulting gene flow.
Long restricted to conservation biology, landscape genetics now benefits from advances in next-generation sequencing, which enable the generation of massive genomic datasets at lower cost and greatly increase sample sizes. This evolution makes it possible to apply these approaches to species of agronomic importance, such as crop pests, which are often characterized by large population sizes, high dispersal capacity, and local dynamics that can be heavily noised by treatments (e.g., Crossley et al., 2017; Parvizi et al., 2024, Caumettes 2025). In this context, landscape genomics offers an integrative framework to link landscape structure, selection pressures, and the spatial diffusion of adaptive variants (Bourguet et al., 2014; Fenderson et al., 2020). The field also benefits from growing access to large, high-resolution datasets made possible by the democratization of remote sensing and satellite imagery, coupled with machine learning methods (Dauphin et al., 2023). These advances allow fine characterization of spatial heterogeneity in agricultural landscapes and analysis of its influence on genetic structure and gene flow. They facilitate the collection of harmonized datasets to conduct multiple landscape-scale studies which are is essential to enhance our understanding of species-environment interactions and to develop effective, context-specific management strategies (Richardson et al. 2016). Inference methods to estimate demographic parameters (population size and dispersal) and selection intensity might also help understand insect pest populations dynamics in agricultural landscapes and the spatio-temporal dynamics of adaptations (e.g., Hancock et al., 2020, Uhl et al. 2024, Caumettes et al. 2025).
This project focuses on a major agricultural pest strongly affected by resistance evolution, the codling moth (Cydia pomonella), which accounts for roughly two-thirds of all insecticide applications in French orchards, making pome fruit one of the most intensively treated crops (Agreste, 2018; Zaller et al., 2023). Most insecticides used against this species face resistance issues, notably the codling moth granulosis virus (CpGV), a biocontrol product widely used against C. pomonella (Ju et al., 2021). We have recently pinpointed the main genes and genetic mechanisms (notably a duplication) underlying resistances to CpGV products, as well as several molecular markers strongly associated with these resistances (Olivares et al., 2023; Idier et al., 2023; Gingueneau et al., in prep).
In this context, this PhD project will generate actionable knowledge for both resistances monitoring and management. It will define the optimal spatial and temporal resolutions relevant for France's Biological Surveillance of the Territory (SBT), specifying the sampling grain required to capture heterogeneity in resistance distribution-both nationally and within production basins-and the optimal monitoring frequency to track temporal dynamics. A particular focus will be placed on identifying production regions at highest risk of widespread CpGV resistance, especially in the PACA region which is the most affected by C. pomonella resistances to CpGV.
The modeling work conducted during this PhD will support the design of collective strategies to manage and prevent resistance spread, in direct collaboration with producer organizations and technical advisors. The thesis is strongly embedded in the PARSADA ASAP program (2025-2030), co-developed with producer organizations, ensuring rapid transfer of results to practitioners and their uptake in coordinated management actions.
By integrating large-scale, multi-source datasets (genomic, landscape, and environmental), this project will deliver a deeper understanding of the drivers of the spatio-temporal dynamics of resistance to a biopesticide that has become central to the management of C. pomonella. The project will deliver a conceptual and methodological framework transferable to other insect pests and resistance issues, thereby strengthening the durability of control strategies.

Chez C. pomonella, des études de génétiques des populations basées sur 7 microsatellites et 24 vergers ont montré une faible structuration génétique neutre à l'échelle de la France et une diversité génétique locale influencée par l'intensité des traitements. En revanche, les marqueurs de résistance aux pyréthrinoïdes (mutation kdr) ont révélé des structures plus marquées que les marqueurs neutres (Franck et al., 2007 ; 2011 ; Voudouris et al., 2012). D'autre part, nos travaux récents ont permis d'importantes avancées dans la caractérisation des bases génétiques des résistances au CpGV, un agent de biocontrôle massivement utilisé contre C. pomonella. Nous avons identifié les principaux gènes et mécanismes génétiques (en particulier une duplication) responsables de ces résistances au CpGV ainsi que plusieurs marqueurs moléculaires fortement associés à ces résistances (Olivares et al. 2023, Idier et al. 2023, Gingueneau et al. in prep). En intégrant des données massives de nature variée (génomiques, paysagères et environnementales), ce projet contribuera à une compréhension approfondie des dynamiques spatio-temporelles des résistances au CpGV, un biopesticide devenu pivot de la lutte contre ce ravageur. L'approche de génomique du paysage permettra de révéler des patrons génétiques fins et identifier les déterminants de la dispersion et de la diffusion des résistances. Ce projet aboutira à l'élaboration d'un cadre conceptuel et méthodologique transférable à d'autres insectes ravageurs et problématiques de résistance, pour renforcer la durabilité des stratégies de lutte.
Understanding the processes that drive the emergence and spread of resistance in agricultural landscapes is critical to preserving the efficacy and durability of control methods. This need is particularly timely in the context of agroecological transition as the reduction of authorized active substances increases reliance on alternative solutions, including biocontrol methods, which are also affected by pest adaptation. Identifying whether populations are resistant or susceptible is a first step, but it is insufficient to predict the increase and spread of resistance within and among production regions, or to identify the levers to contain it. Spatio-temporal dynamics of resistance depend not only on the variability of local selection pressures but also on demographic processes governing pest populations (Lenormand, 2002; Miller, 2017). Population size and dispersal depend on the components of the environmental matrix, through biotic features (e.g., host plants, landscape structure) and abiotic drivers (e.g., farming practices, climate). In agricultural landscapes, this matrix can be highly dynamic in space and time due to the diversity, abundance, distribution, and phenology of host plants, as well as the diversity of farming practices. It is therefore essential to quantify functional connectivity by clarifying how the environmental matrix modulates dispersal and the resulting gene flow.
Long restricted to conservation biology, landscape genetics now benefits from advances in next-generation sequencing, which enable the generation of massive genomic datasets at lower cost and greatly increase sample sizes. This evolution makes it possible to apply these approaches to species of agronomic importance, such as crop pests, which are often characterized by large population sizes, high dispersal capacity, and local dynamics that can be heavily noised by treatments (e.g., Crossley et al., 2017; Parvizi et al., 2024, Caumettes 2025). In this context, landscape genomics offers an integrative framework to link landscape structure, selection pressures, and the spatial diffusion of adaptive variants (Bourguet et al., 2014; Fenderson et al., 2020). The field also benefits from growing access to large, high-resolution datasets made possible by the democratization of remote sensing and satellite imagery, coupled with machine learning methods (Dauphin et al., 2023). These advances allow fine characterization of spatial heterogeneity in agricultural landscapes and analysis of its influence on genetic structure and gene flow. They facilitate the collection of harmonized datasets to conduct multiple landscape-scale studies which are is essential to enhance our understanding of species-environment interactions and to develop effective, context-specific management strategies (Richardson et al. 2016). Inference methods to estimate demographic parameters (population size and dispersal) and selection intensity might also help understand insect pest populations dynamics in agricultural landscapes and the spatio-temporal dynamics of adaptations (e.g., Hancock et al., 2020, Uhl et al. 2024, Caumettes et al. 2025).
This project focuses on a major agricultural pest strongly affected by resistance evolution, the codling moth (Cydia pomonella), which accounts for roughly two-thirds of all insecticide applications in French orchards, making pome fruit one of the most intensively treated crops (Agreste, 2018; Zaller et al., 2023). Most insecticides used against this species face resistance issues, notably the codling moth granulosis virus (CpGV), a biocontrol product widely used against C. pomonella (Ju et al., 2021). We have recently pinpointed the main genes and genetic mechanisms (notably a duplication) underlying resistances to CpGV products, as well as several molecular markers strongly associated with these resistances (Olivares et al., 2023; Idier et al., 2023; Gingueneau et al., in prep).
In this context, this PhD project will generate actionable knowledge for both resistances monitoring and management. It will define the optimal spatial and temporal resolutions relevant for France's Biological Surveillance of the Territory (SBT), specifying the sampling grain required to capture heterogeneity in resistance distribution-both nationally and within production basins-and the optimal monitoring frequency to track temporal dynamics. A particular focus will be placed on identifying production regions at highest risk of widespread CpGV resistance, especially in the PACA region which is the most affected by C. pomonella resistances to CpGV.
The modeling work conducted during this PhD will support the design of collective strategies to manage and prevent resistance spread, in direct collaboration with producer organizations and technical advisors. The thesis is strongly embedded in the PARSADA ASAP program (2025-2030), co-developed with producer organizations, ensuring rapid transfer of results to practitioners and their uptake in coordinated management actions.
By integrating large-scale, multi-source datasets (genomic, landscape, and environmental), this project will deliver a deeper understanding of the drivers of the spatio-temporal dynamics of resistance to a biopesticide that has become central to the management of C. pomonella. The project will deliver a conceptual and methodological framework transferable to other insect pests and resistance issues, thereby strengthening the durability of control strategies.
This thesis adopts a resolutely interdisciplinary approach, bridging remote sensing (providing high-resolution data on orchard distribution and habitat heterogeneity, thus enabling precise quantification of environmental selection pressures), landscape genomics (deciphering neutral and adaptive genetic structure, thus revealing patterns of gene flow, dispersal, and local adaptation in C. pomonella populations), spatial modelling (synthesising environmental, genomic, and treatment data to simulate resistance spread and evaluate the impact of management scenarios), and machine learning (uncovering non-linear relationships and key drivers of resistance evolution, thus enhancing predictive accuracy) to unravel the complex dynamics of CpGV resistance at the landscape scale. The project is supported both by the composition of the supervision team (1 landscape genomics specialist, 1 statistical modeler, 1 bioinformatician) and its collaboration with a consortium of experts in these fields, ensuring a comprehensive, multi-scale analysis from molecular mechanisms to regional management strategies delivering practical solutions for sustainable pest control.