Thèse un Workflow In Silico-In Vitro pour Identifier et Valider des Variants Génomiques Utiles Contre les Maladies Infectieuses Intestinales chez le Porc. H/F - 75

Établissement : Université Paris-Saclay GS Life Sciences and Health
École doctorale : Structure et Dynamique des Systèmes Vivants
Laboratoire de recherche : GABI - Génétique animale et Biologie intégrative
Direction de la thèse : Elisabetta GIUFFRA ORCID 0000000195682056
Début de la thèse : 2026-10-01
Date limite de candidature : 2026-07-31T23:59:59

Dans le contexte du changement climatique, il est essentiel de prévenir les infections pathogènes et d'anticiper les épidémies infectieuses, non seulement pour l'industrie porcine, mais aussi dans l'optique des programmes One Health et d'EcoHealth. La sélection tend à réduire la diversité génétique des populations. Une conséquence potentielle est la perte d'allèles ou variants génétiques clés des populations ce qui limite leur capacité à s'adapter aux pathogènes récurrents et émergents. On peut supposer que les programmes d'élevage permettant de maintenir la diversité génétique au niveau des loci clés de la réponse immunitaire de l'hôte pourraient offrir aux éleveurs de porcs un niveau de protection durable. D'autre part, la collecte des caractères de résistance à des maladies infectieuses in vivo est onéreuse et éthiquement discutable. Dans ce contexte des 3R du bien-être animal en recherche (remplacement, réduction, raffinement), le phénotypage in vitro offre une alternative. L'objectif de la thèse est d'établir un workflow in silico-in vitro permettant d'identifier la diversité génétique de loci sélectionnés et valider son caractère fonctionnel, dans des populations commerciales. Bien que motivé par l'objectif à long terme de fournir des données pour modéliser la préservation de leur diversité tout en maintenant les objectifs de production, ce sujet dépasse le cadre de la thèse.
Pour sa réalisation, le projet s'appuie sur l'ensemble des connaissances disponibles sur les récepteurs viraux et les facteurs de restriction de l'hôte sur le SarsCov-2 humain et d'autres coronavirus, y compris les coronavirus entériques porcins (TGEV, virus de la gastro-entérite transmissible, et PEDV, virus de la diarrhée épidémique porcine). Ces derniers sont connus pour causer de lourdes pertes dans les productions porcines, soit directement, soit en prédisposant à des infections secondaires. Nous avons ainsi identifié 150 gènes hôtes clés, correspondant à des récepteurs et à des facteurs de restriction pan-viraux identifiés dans des études antérieures sur les coronavirus. Leurs variants génétiques représentent des facteurs de résistance potentiels aux infections pathogènes.
Sur la base des variants génétiques localisées dans ces gènes et identifiées dans les troupeaux commerciaux ciblés, le doctorant devra : i) développer et utiliser des approches permettant de prédire l'impact fonctionnel des variants codants et régulateurs sur l'expression des gènes et la fonction des protéines ; ii) valider expérimentalement les variants prédits comme les plus impactants dans des modèles de lignées cellulaires, et caractériser leurs conséquences phénotypiques lors d'une infection par le TGEV à l'aide de systèmes hétérologues et organoïdes in vitro.
Le présent projet ouvre la voie à une gestion efficace et hautement flexible de la biodiversité fonctionnelle des populations porcines commerciales, fondée sur des approches génomiques. Cette dernière est conforme au principe des 3R , adaptée à tout locus économiquement pertinent, et pouvant être étendue à d'autres espèces d'élevage.

Global changes favour recurrent and emerging pathogens to spread rapidly among farm animals, with evidence of different levels of genetic resistance and susceptibility (Knap, Doeschl-Wilson, 2020). The 1992 Convention on Biological Diversity established genetic diversity as a reservoir for breeding, particularly to combat infectious diseases. Genomic selection breeding schemes used to improve production traits can lead to a reduction of the genetic variability of commercial populations, thus biodiversity preservation is a key part of large concerted efforts to build more resilient and sustainable pig populations (The conservation of populations, 2008). While the characterization of genetic diversity has traditionally been focused on neutral diversity, the improved functional knowledge of animal genomes allows to formulate strategies aiming to preserve diversity at target loci of interest. The interest of this approach has been shown in the context of wild species conservation (Texeira 2021) and has been noticed in sheep and animal models for the MHC genes (Corwall et al 2018).
Preserving genetic diversity at key loci undermining immune responses to infectious diseases (and not focusing solely on neutral variations, as classically done) could be a durable, cost effective and potentially sustainable approach with ongoing efforts to re-establish the commercial population diversity upon infection breakout. Recurrent and emerging viruses are particularly relevant to pig production due to their strong economic impact. The recent African Swine Fever outbreak demonstrated the vulnerability of pig production: in China alone, the pig population was reduced by a third. Indeed, despite strict biosecurity measures, infections remain a significant challenge for pig breeders because vaccines are expensive and time consuming to administer and treatments are rarely effective at early age. Health traits are a major issue for the pig chain, thus a major competitive key to face changing and alternative breeding conditions. However, selecting more resistant or resilient animals is challenging because defining easily measurable and affordable characteristics is difficult. Only few successful stories are reported, as for bovine tuberculosis (Banos G, 2023). The nucleus populations selected in pigs are housed in limited pathogen-pressure conditions making it difficult to elaborate genomic prediction tools for herd vulnerability.
Disease traits are typically complex, as they are the product of (often) multi-pathogen infections and of several other environmental factors. However, particularly with viral pathogens, it has been shown that even regulatory variants at specific loci could explain considerable proportions of the variance of immune responses (e.g., PCV2 (Opriessnig et al 2006; Liu et al 2025). The Sars-Cov2 pandemic has fuelled research on Coronaviruses and several CRISPR genome-wide screens have identified dozens of pan-virus restriction factors influencing viral entry in host cells from different species, including pigs (Beumer et al 2021; Kruse et al 2021, Lin et al 2025). The major porcine Coronavirus species are TGEV and PEDV, both causing fatal diarrhoeal epidemics in piglets under 2 weeks of age and easily transmitted in the context of intensive farming (Mclean et al 2022).
Understanding genome function is essential for identifying the genetic variants underlying phenotypes of interest in livestock, including resistance to pathogens. Large-scale initiatives such as the Functional Annotation of Animal Genomes (FAANG) and FarmGTEx projects provide valuable resources to characterize functional genomic elements and improve our understanding of genotype-phenotype relationships. Building on these resources, the development of reliable variant effect prediction methods based on deep learning, combined with functional assays in cellular models, offers the opportunity to efficiently screen large numbers of variants and validate their effects experimentally. This integrative strategy, well established in human genetics, has proven effective in refining the interpretation of regulatory and coding variants and accelerating the identification of causal variants (PMID: 39874404). More recently, similar approaches have gained increasing interest in livestock species, as they provide powerful tools to dissect the molecular mechanisms underlying complex traits and disease resistance. Such strategies allow to move from association signals to causal variants with direct biological relevance, as illustrated by our recent work (Charles et al., 2025).
Organoids have emerged as powerful experimental models that can serve as proxies for in vivo tissues (reviewed by Zhao et al 2022). Over the past fifteen years, major advances in stem cell biology have enabled the long-term cultivation of tissue-derived adult stem cells and pluripotent stem cells, leading to the generation of self-organizing three-dimensional organoids. These structures reproduce key cellular compositions, spatial organization, and functional properties of their tissue of origin. In the context of infectious diseases, organoids provide physiologically relevant in vitro models to study viral infection and host-pathogen interactions. Their three-dimensional organization and cellular heterogeneity enable analysis of viral entry routes, replication cycles, cell-type specificity, and tissue-level responses. Li et al 2020 showed TGEV infection of apical-out porcine gut organoid.

This thesis project has two main objectives: 1) to predict cis-regulatory and coding variants at virus receptors and pan-virus restriction factors in the pig genome; 2) to functionally validate the effects of top variants that segregate in porcine commercial herds using heterologous systems and organoids upon infection with TGEV in vitro.

WP 1: Prediction and hierarchisation of regulatory and coding variants affecting TGEV and PEDV receptors and pan-virus restriction factors.
In order to explore the natural variation interesting to maintain in herds to diversify targets of infections, we first needed to define genes known to be used for entry and post-entry by Coronaviruses. Based on the available data from genome-wide CRISPR screens and other functional studies in Coronavirus, we have assembled a preliminary list of 150 genes proved to be involved in the entry and post-entry processes of most Coronaviruses in humans, pigs and other species (e.g. ACE2, ANPEP,. DPP4, IFITM1).
Project partners will supply imputed sequences covering the 150 loci in the porcine population considered under investigation. We will identify cis-regulatory and splicing variants at the 150 gene loci by relying on the above output and in parallel on open-source databases such as the PigGTex portal (https://piggtex.farmgtex.org). Several tools for variant annotation and functional prediction are in use at GABI. The pipelines for predicting the impact of regulatory and coding variants have been established (SnpEff, VEP, as well as Splice AI and Pangolin, AlphaFold, CADD and internal tools like STIP) and most of them have been adapted or developed by members of GABI. The candidate will have the possibility to apply and further update these and related available pipelines. The variants predicted to impact on gene expression will be further prioritised based on their association with the cis-eQTLs (expression quantitative trait loci) previously identified in the LW animals by the european project GENE-SWitCH and with the eQTLs and s-QTLs (splicing QTLs) compiled in the PigGTex portal. Haplotypes, defined as a set of variants positioned in the same chromosome, and inherited together from a single parent, will be reconstructed in the population data.The accuracy of these predictions will be further ascertained by using available RNAseq data from gut and lung (main target tissues of Coronaviruses).
WP2: Functional validation of top variants in vitro with TGEV virus
Task 2.1: To assess the effects of top regulatory variants in heterologous systems To assess regulatory variants that impact transcription rate, we will clone regulatory regions containing either the reference or alternative alleles into luciferase reporter constructs. To assess the effect of variants on splicing, low-throughput tests using Full Length Gene Assay (FLGA) (Gaiani et al., 2023) or minigenes (Boulling et al., 2025) set up at GABI will be used for in vitro validation. pcDNA3.1 and pcAT7-Glo1 expression vectors will be used to produce full length gene constructs or minigenes, respectively. Luciferase as well as expression constructs will be transfected into heterologous systems such as HEK293 or porcine epithelial PK15 cell lines to quantify promoter/enhancer activity or splicing efficiency. These cells present a high transfection efficiency, enabling reliable measurement of reporter activity, protein expression or splicing. RNA will be extracted and amplified by RT-PCR to analyse the splicing pattern of reference and alternative alleles to identify functional variants.
Task 2.2: To determine the relevance of susceptibility top coding variants for fusion to the TGEV Spike protein. Among all coronavirus proteins, the Spike is critical. Indeed, entry of porcine coronaviruses is strictly dependent on their Spike protein, which functions 1) by binding to a specific receptor present on the surface of its target cell, and 2) by inducing fusion of cell and viral membranes. The Spike is also the main target of neutralizing antibodies elicited after natural infection or vaccination, as illustrated during the Covid-19 pandemic during which efficient vaccines were based on the Spike mRNA sequence. Our team developed a heterologous expression system for the Spike proteins of various human and porcine coronaviruses, to screen and quantitatively study the effect of host proteins/molecules that facilitate or block this fusion activity (Buchrieser et al., 2019), see attached pdf: Fig. 1A, B). Preliminary results indicate that co-expression of SARSCoV-2 Spike with its receptor ACE2 in this GFP-Split heterologous system induces a massive fusion, not observed in the presence of C2, a Spike-ACE2 binding inhibitor (see attached pdf:Fig. 1C). Cells will be transfected by plasmids encoding the Spike protein of TGEV and simultaneously transfected with plasmids encoding candidates with the identified coding SNPs before being challenged with the virus. Identical assays will be performed with PEDV spike protein.
Task 3 Validation of haplotypes predicted to affect ANPEP expression vs. the individual genetic backgrounds The variability in innate immune response genes contributes to differing levels of susceptibility to viral infection. While knockout (ko) of the ANPEP gene is known to prevent TGEV infection (Whitworth et al. 2019), the effect on viral infection of regulatory mutations at this locus is unknown. We have biobanked organoids from LW preweaning and post-weaning animals for which ANPEP ko have been produced with conventional Cas9 gene inactivation (see attached pdf:Figure 1D). After sequencing of the endogenous loci in the biobanked organoids, we will repeat genome editing to introduce the haplotype variants identified in T2 and we will quantitatively assess the capacity of these organoids to get infected with TGEV or PEDV using wt and ko organoids as controls.