Soybeans are an economically important plant, with an annual crop value that consistently exceeds 20 billion dollars in the United States alone. A recent increase in demand for soybeans, stemming from its diverse applications in products such as animal feed, oil, and biofuel, has created an emphasis for soybean breeders in value added cultivars. These cultivars, have improved, or altered, agronomic or seed composition traits, allowing them to be efficiently utilized in a specific niche of the processing industry. Facilitating the development of such cultivars requires a thorough understanding of the genetic factors that affect the manifestation of value added traits. Value added traits investigated in this study include seed sucrose, raffinose, stachyose, and phytate content, seed weight, and maturity. The objective of the first part of this project was to characterize the source of low seed stachyose in soybean line PI200508. Two F2 populations, developed from PI200508 and soybean introductions which exhibited higher seed stachyose content were utilized in a QTL analysis approach that incorporated the use of the Williams82 whole genome shotgun (WGS) sequence (http://www.phytozome.org) in a candidate gene mapping approach. A predicted soybean galactosyltransferase gene was established as a candidate gene due to its observed segregation with the single low stachyose QTL observed on molecular linkage group (MLG) C2 in both populations. Sequencing of this putative gene revealed a unique 3 bp deletion in PI200508. A marker developed to exploit this deletion accounted for 88% and 94% of the phenotypic variance for seed stachyose content in the two experimental populations, highlighting its potential for use in marker assisted selection of the PI200508 source of low raffinose and stachyose. The second part of this project involved QTL analysis of seed sucrose, raffinose, stachyose, and phytate content, as well as seed weight in a linkage map for a F8 RIL population developed from the Glycine max line V71-370 and the Glycine soja introduction PI40712. Analysis across all 20 soybean MLG identified 25 QTL for these traits on MLG A1, A2, C2, D1b, D2, F, G, H, I, L, M, O. Nine of these QTL were supported across multiple environments, indicating that they, and their associated markers, could be useful to breeders working with these traits. The third part of this project used the same F8 RIL linkage map to investigate time to maturity (Reproductive stage R8). V71-370 and PI407162 differ in time to maturity when grown in Virginia, and the RILs developed from this cross displayed a wide range in maturity. Two major QTL were identified on MLG H and L. Examination of the Williams82 WGS sequence in these QTL regions revealed two predicted genes with homology to Arabidopsis thaliana light response and photoperiodism genes which were investigated as candidate soybean maturity genes. Markers developed from these predicted genes showed close association with the observed QTL, and could facilitate the further investigation of this complex trait. / Ph. D.

Fusarium head blight (FHB), caused by Fusarium graminearum Schwabe, is a devastating fungal disease that affects barley production worldwide through damaged kernels, low yields, and production of mycotoxins. The most effective means of control is through the use of FHB resistant cultivars that are developed through gene pyramiding and incorporation of various sources of resistance. Resistance identified in winter barley cultivar Eve has been identified in Virginia Tech's Small Grains Breeding Program. The objectives of this study are to validate resistance and identify additional resistance QTL, and identify tightly linked and diagnostic markers for use in MAS programs. A population of F5:7 RILs derived from the cross Eve/Doyce were evaluated for FHB resistance. FHB incidence (Inc), FHB severity (Sev), Fusarium damaged kernels (FDK), and deoxynivalenol (DON) were assessed over 2014-15 and 2015-16 at Blacksburg, VA, Mount Holly, VA, Kinston, NC, and Lexington, KY. A QTL on chromosome 6H associated with FHB Sev, FDK, and DON was identified. The QTL accounted for as high as 14.5% of the phenotypic variation for DON accumulation, 8.3% for FHB Sev, and 11.87% to 17.63% for FDK. The QTL marker region spans 56.5 to 66.6 cM with flanking markers SCRI_RS_147342 and Bmag0613. QTL for morphological traits, heading date and height, were found in the same region as the FHB resistance traits. Diagnostic SNP makers associated with the FHB resistance QTL identified can be used for FHB resistance identification in MAS breeding programs to incorporate and pyramid resistance QTL and genes into other plant material. / Master of Science

Fusarium Head Blight (FHB) is a devastating fungal disease of wheat (Triticum aestivum L.) and results in significant economic losses due to reductions in grain yield and the accumulation of mycotoxins, such as deoxynivalenol (DON) and nivalenol (NIV). As a result, breeding programs have been working to identify resistance genes in wheat varieties known to be resistant to FHB. Some of the major quantitative trait loci (QTL) for FHB resistance identified to date have been from exotic sources such as 'Sumai3' and the Chinese landrace Wangshuibai, and native resistance has been identified in North American cultivars such as Ernie and Truman which are being used in breeding programs. This study was conducted to characterize and map QTL for resistance to FHB in the soft red winter wheat cultivar Jamestown and to identify tightly linked DNA markers associated with those QTL so that marker-assisted selection (MAS) can be used in pyramiding these and other known QTL into elite backgrounds. Types of resistance assessed in this study include: Type I (resistance to initial infection; incidence), Type II (resistance to spread in wheat spike; severity), and decreases in mycotoxin accumulation (DON) and percentage of Fusarium damaged kernels (FDK). A population composed of 186 F5:7 recombinant-inbred lines (RILs) from the cross Pioneer Brand '25R47'/Jamestown were used to evaluate these traits in six environments (MD, NC, and VA in 2011 and 2012). This study identified a QTL for resistance to DON accumulation and FHB severity on the wheat chromosome 1B. The QTL accounted for 12.7% to 13.3% of the phenotypic variation in DON accumulation and 26.1% of the phenotypic variation in FHB severity. The most diagnostic marker for the QTL on chromosome 1B associated with resistance to FHB severity and DON accumulation is Xwmc500.6 located 7.2 cM from the QTL peak and flanked by markers Xwmc500.7 and Xgwm273.2 (28.2 cM interval). / Master of Science

Dans le contexte de raréfaction de la ressource en eau, il est nécessaire de prendre en compte le caractère de tolérance à la sécheresse dans les programmes d’amélioration variétale, en particulier pour les porte-greffes de vigne. L’objectif de cette thèse est d’analyser, à l’échelle de la plante entière, le déterminisme génétique des effets du porte-greffe sur les réponses de la tranpiration et de la croissance, au niveau de la partie aérienne, en conditions de déficit hydrique édaphique. L’étude de ces caractères complexes, soumis à une forte interaction génotype × environnement, a été conduite à partir d’une approche de génétique quantitative où ces interactions ont été intégrées par deux approches : l’utilisation des paramètres de courbes de réponses comme caractère quantitatif et la mise en oeuvre d’une analyse QTL pluriannuelle. Une population de type F1, issu du croisement Vitis vinifera Cabernet Sauvignon × Vitis riparia Gloire de Montpellier, constituée de 138 individus, a été cartographiée génétiquement avec des marqueurs microsatellites. Cette descendance a été ensuite phénotypée en position de porte-greffe, avec un greffon unique. A partir de cet outil, des QTL ont été recherchés et identifiés pour la transpiration, l’efficience d’utilisation de l’eau et la plasticité de la transpiration induite par le porte-greffe, ainsi que pour plusieurs paramètres de la croissance et de l’allocation de la biomasse au sein de la plante entière. La plasticité de la transpiration réponse à l’état hydrique du substrat est déterminée génétiquement par des régions distinctes des autres caractères descriptifs de la consommation en eau. La colocalisation de gènes candidats dans les intervalles de confiance des QTL, impliqués dans la signalisation chimique (ABA) ou hydraulique (aquaporines) permet de formuler des hypothèses sur la signalisation porte-greffe/ greffon en condition de déficit hydrique. Des régions spécifiques du génome paraissent contrôler la croissance et la vigueur conférée par le porte-greffe en conditions de déficit hydrique. Ces régions sont également indépendantes de celles contrôlant la transpiration ou l’efficience d’utilisation de l’eau. Ce résultat permet d’envisager un travail de sélection pour ces deux catégories de caractères de manière indépendante / In the water scarcity context with the global climate change, drought tolerance must be taken into account in crop genetic improvement program, particularly for grapevine rootstocks. The objective of this thesis was to analyse at the whole plant level, the genetic determinism of rootstock effect on the transpiration and growth of the scion, under edaphic water deficit conditions. The study of these complex traits, submitted to a strong genotype × environment interaction, was performed with a quantitative genetic approach. Genotype × environment interactions were integrated with two methods: first, using response curve to an environmental variable for detecting QTL, and secondly, to combine data in a multi-environment QTL analysis. The pedigree population consisted of 138 F1 individuals derived from the interspecific cross of Vitis vinifera Cabernet Sauvignon × Vitis riparia Gloire de Montpellier. This family was mapped with single sequence repeats (SSR) markers. This population was assessed as rootstock, so every genotype was grafted with the same scion. Then, QTL were detected for transpiration, water use efficiency and transpiration plasticity induced by rootstock, and for growth and biomass allocation inside the whole plant. Transpiration plasticity was represented as a function of substrate water status and was genetically determined with distinct genome regions from the other traits related to water consumption. Candidate genes involved in hormonal (ABA) or hydraulic (aquaporins) signaling between rootstock and scion, under water deficit conditions, were localized in the QTL confidence interval. Some specific genome regions were involved in growth and confered vigour genetic determinism under water deficit conditions. These regions were also different from those identified for transpiration or water use efficiency. This result allows a further selection process for these two traits groups independently.

Des génotypes adaptés à des environnements contrastés ont de grandes chances de se comporter différemment lorsqu’ils sont placés dans des conditions similaires et contrôlées, notamment si leur sensibilité aux signaux environnementaux et/ou leur croissance intrinsèque sont limitées à différents niveaux. De ce fait, la variabilité observée dans les populations naturelles peut être utilisée comme une source illimitée de nouveaux allèles ou gènes pour l’étude des bases génétiques de la variation des traits quantitatifs. Mon travail de doctorat a consisté en l’analyse de la variabilité naturelle pour la croissance et la réponse à l’environnement chez Arabidopsis thaliana. Le but des approches de génétique quantitative est de comprendre comment la diversité génétique et épigénétique contrôle la variabilité phénotypique observée dans les populations à différentes échelles, au cours du développement et sous différentes contraintes environnementales. De plus, ces analyses ont pour objectif de comprendre comment les processus adaptatifs et démographiques influencent la fréquence de ces variants dans les populations en fonction de leur environnement local. Ainsi, l’étude de la variabilité naturelle peut être appréhendée en utilisant diverses approches, de la génétique et des méthodes de biologie moléculaire aux études écologiques et évolutives. Au cours de mon doctorat, j’ai eu la chance de travailler sur plusieurs de ces aspects au travers de trois projets indépendants qui exploitent tous la variabilité naturelle d’A. thaliana.Le premier projet a consisté en l’analyse du pattern de polymorphisme observé dans des populations d’A. thaliana au gène MOT1 qui code pour un transporteur de molybdate (la forme assimilable du molybdène (Mo), un micro-élément essentiel) et qui est responsable d’une partie des variations de croissance et de fitness observées à l’échelle de l’espèce en fonction de la disponibilité en Mo des sols. J’ai montré à différentes échelles géographiques que le pattern de polymorphisme à MOT1 ne reflète pas une évolution neutre mais présente plutôt des traces de sélection diversifiante. Ce travail a contribué à renforcer l’hypothèse selon laquelle des mutations au niveau du gène MOT1 pourraient avoir été sélectionnées dans certaines populations pour faire face aux niveaux élevés de Mo observés dans certains sols et potentiellement délétères malgré leur effet négatif sur des milieux pauvres en Mo.Le deuxième projet portait sur la caractérisation et l’analyse fonctionnelle de deux récepteur-kinase putatifs (RLK) identifiés de part leurs effets sur la croissance foliaire spécifiquement en réponse à un stress induit par du mannitol mais pas sous d’autres contraintes osmotiques. La fonction de ces récepteurs chez A. thaliana -qui n’est pas connu pour produire du mannitol- peut paraître intrigante. Les différentes expériences réalisées au cours de cette thèse nous ont cependant permis de construire un modèle selon lequel ces récepteurs pourraient être activés par le mannitol produit par certains pathogènes tel que les champignons et participer aux réponses de défense de la plante.Le troisième projet a été réalisé en collaboration avec l’équipe de Michel Vincentz (CBMEG, Brésil) et de Vincent Colot (IBENS, Paris) et consiste en l’analyse de l’occurrence de variants épigénétiques naturels au gène QQS dans différentes populations d’Asie Centrale et de leurs possibles conséquences phénotypique et adaptative.En conclusion, l’analyse des variants génétiques et épigénétiques naturels à l’origine des variations de biomasse en interaction avec l’environnement permet de comprendre comment l’évolution façonne la variabilité naturelle. / Genotypes adapted to contrasting environments are expected to behave differently when placed in common controlled conditions, if their sensitivity to environmental cues or intrinsic growth behaviour are set to different thresholds, or are limited at distinct levels. This allows natural variation to be exploited as an unlimited source of new alleles or genes for the study of the genetic basis of quantitative trait variation. My doctoral work focuses on analysing natural variation for shoot growth and response to the environment in A. thaliana. Natural variation analyses aim at understanding how molecular genetic or epigenetic diversity controls phenotypic variation at different scales and times of plant development and under different environmental conditions, and how selection or demographic processes influence the frequency of those molecular variants in populations for them to get adapted to their local environment. As such, the analysis of A. thaliana natural variation can be addressed using a variety of approaches, from genetics and molecular methods to ecology and evolutionary questions. During my PhD, I got the chance to tackle several of those aspects through my contributions to three independent projects which have in common to exploit A. thaliana natural variation. The first one is the analysis of the pattern of polymorphism from a set of 102 A. thaliana accessions at the MOT1 gene coding for a molybdate transporter (an essential micronutrient) and responsible for contrasted growth and fitness among accessions in response to Mo availability in the soil. I showed at different geographical scales that MOT1 pattern of polymorphisms is not consistent with neutral evolution and shows signs of diversifying selection. This work helped reinforce the hypothesis that in some populations, mutations in MOT1 have been selected to face soils rich in Mo and potentially deleterious despite their negative effect on Mo-limiting soils. The second project consists in the characterisation and functional analysis of two putative receptor-like kinases (RLKs) identified from their effect on shoot growth specifically under mannitol-supplemented media and not in response to other osmotic constraints. The function of such RLKs in A. thaliana, which is not known to synthesize mannitol was intriguing at first but, through different experiments, we built the hypothesis that those RLKs could be activated by the mannitol produced by some pathogens such as fungi and participate to plant defensive response. The third project, in collaboration with Michel Vincentz’s team from CBMEG (Brasil) and Vincent Colot (IBENS, Paris), consists in the analysis of the occurrence of natural epigenetic variants of the QQS gene in different populations from Central Asia and their possible phenotypic and adaptive consequences. Overall, these analyses of the genetic and epigenetic molecular variation leading to the biomass phenotype(s) in interaction with the environment provide clues as to how and where in the pathways adaptation is shaping natural variation.

L’optimisation de la biomasse produite par les taillis à courte rotation de peupliers représente un enjeu majeur pour la production de biocarburants de deuxième génération. Dans ce contexte, ce travail vise à faciliter le développement à court terme de nouvelles variétés clonales de peuplier permettant la production d’une ressource de qualité en s’intéressant plus particulièrement à l’architecture génétique de la résistance à la rouille foliaire et de la qualité du bois chez le peuplier noir (Populus nigra), espèce parente des hybrides cultivés. Des marqueurs SNP ont été développés à partir du séquençage de 665 fragments de gènes dans un panel de 21 individus. Ces derniers ont été associés à des marqueurs SSR et AFLP pour construire de nouvelles cartes génétiques sur une famille de 324 plein-frères clonés. Une technique de phénotypage à haut débit basée sur la spectrométrie à proche infrarouge a été développée pour prédire les teneurs en composés chimiques du bois ainsi que le rendement en saccharification. La mise en évidence d’une variabilité génétique importante pour l’ensemble des caractères a permis de cartographier les régions génomiques impliquées dans leur variation. Parmi les 11 QTL détectés pour la résistance, un QTL à effet majeur co-localise avec un QTL majeur associé à la résistance à la rouille foliaire du saule. Pour la qualité du bois, 15 QTL à effet faible à moyen ont été détectés, dont un cartographié sur le chromosome XIII qui colocalise avec des QTL précédemment identifiés chez le peuplier pour les teneurs en sucres et en lignines. Ce travail de thèse ouvre des perspectives d’identification de gènes sous-jacents aux QTL par génétique d’association. / Improvement of lignocellulosic resources from poplar short rotation coppices is a major challenge for the production of second generation biofuels. In this context, the present work aims at optimizing short term creation and deployment of improved poplar clonal varieties through the dissection of genetic control of both leaf rust resistance and wood quality in black poplar (Populus nigra), one of the parental species of cultivated hybrids. SNP markers have been developed from the resequencing of 665 gene fragments in a discovery panel of 21 individuals. These markers were combined with SSRs and AFLPs to build new genetic maps in a pedigree composed of 324 cloned full-sibs. High throughput phenotyping based on near infrared spectroscopy has been used to predict wood chemical contents and saccharification yield. High genetic variability expressed in all traits allowed the identification of genomic regions controlling this variation. Of the 11 QTL mapped for resistance, one major QTL co-localized with a major QTL previously detected for leaf rust resistance in willow. For wood quality, 15 QTL with low to moderate effect have been identified. Interestingly, one QTL mapped on chromosome XIII and colocalized with sugar and lignin contents QTL previously detected in poplar. Present results open perspectives towards the identification of candidate genes underlying the detected QTL through association genetics.

We propose Bayesian methods for selecting and estimating different types of mixture models which are widely used inGenetics and MolecularBiology. We specifically propose data-driven selection and estimation methods for a generalized mixture model, which accommodates the usual (independent) and the first-order (dependent) models in one framework, and QTL (quantitativetrait locus) mapping models for independent and pedigree data. For clustering genes through a mixture model, we propose three nonparametric Bayesian methods: a marginal nested Dirichlet process (NDP), which is able to cluster distributions and, a predictive recursion clustering scheme (PRC) and a subset nonparametric Bayesian (SNOB) clustering algorithm for clustering bigdata. We analyze and compare the performance of the proposed methods and traditional procedures of selection, estimation and clustering in simulated and real datasets. The proposed methods are more flexible, improve the convergence of the algorithms and provide more accurate estimates in many situations. In addition, we propose methods for estimating non observable QTLs genotypes and missing parents and improve the Mendelian probability of inheritance of nonfounder genotype using conditional independence structures.We also suggest applying diagnostic measures to check the goodness of fit of QTLmappingmodels. / Nos propomos métodos Bayesianos para selecionar e estimar diferentes tipos de modelos de mistura que são amplamente utilizados em Genética e Biologia Molecular. Especificamente, propomos métodos direcionados pelos dados para selecionar e estimar um modelo de mistura generalizado, que descreve o modelo de mistura usual (independente) e o de primeira ordem numa mesma estrutura, e modelos de mapeamento de QTL com dados independentes e familiares. Para agrupar genes através de modelos de mistura, nos propomos três métodos Bayesianos não-paramétricos: o processo de Dirichlet aninhado que possibilita agrupamento de distribuições e, um algoritmo preditivo recursivo e outro Bayesiano não- paramétrico exato para agrupar dados de alta dimensão. Analisamos e comparamos o desempenho dos métodos propostos e dos procedimentos tradicionais de seleção e estimação de modelos e agrupamento de dados em conjuntos de dados simulados e reais. Os métodos propostos são mais flexíveis, aprimoram a convergência dos algoritmos e apresentam estimativas mais precisas em muitas situações. Além disso, nos propomos procedimentos para estimar o genótipo não observável dos QTL se de pais faltantes e melhorar a probabilidade Mendeliana de herança genética do genótipo dos descendentes através da estrutura condicional de independência entre as variáveis. Também sugerimos aplicar medidas de diagnóstico para verificar a qualidade do ajuste dos modelos de mapeamento de QTLs.

Le feu bactérien et le psylle causent d’importantes pertes économiques dans les zones de production du poirier dans le monde entier. Le développement de nouvelles variétés de poirier résistantes à ces bio-agresseurs constitue un enjeu majeur dans le cadre d’un programme de lutte intégrée. L’objectif de ce projet de thèse est l'étude du déterminisme génétique de la résistance vis-à-vis de ces deux bio-agresseurs. La thèse a été réalisée dans le cadre d’une collaboration internationale entre Fondazione Edmund Mach (Italie), Institut de Recherches en Horticulture et Semences (France) et Plant & Food Research (Nouvelle-Zélande). Une descendance interspécifique de poirier T003 x ‘Moonglow’ a été développée avec pour objectif de cumuler les résistances au feu bactérien et au psylle provenant de variétés asiatiques et européennes de Pyrus. Deux cartes génétiques ont été élaborées pour T003 et ‘Moonglow’ sur la base de marqueurs SNP (Single Nucleotide Polymorphism) et SSR (microsatellite), et la cartographie de QTLs (Quantitative Trait Loci) a permis de démontrer le déterminisme polygénique de la résistance à ces bio-agresseurs. Une sélection assistée par marqueurs (MAS) peut donc être engagée pour ces deux caractères. Des incompatibilités génétiques ont aussi été observées dans une partie de la descendance, ce qui a permis de cartographier pour la première fois chez le poirier les zones du génome liées au phénomène de « nécrose hybride ». Le développement de marqueurs liés aux gènes létaux devrait permettre aux sélectionneurs d’éviter les combinaisons incompatibles en croisement qui peuvent impacter certains caractères agronomiques co-ségrégant avec ces gènes létaux. / The goal of this PhD project was to study the genetic architecture of pear resistance to two of its most significant diseases and pests, fire blight and psylla, which cause severe yield losses in all the main pear production regions worldwide. The development of new pear varieties with resistance against these two biotic stresses is of major interest for Integrated Pest Management. This project was designed in a joint collaboration among Fondazione Edmund Mach (Italy), Institut de Recherches en Horticulture et Semences (France) and Plant & Food Research (New-Zealand). The interspecific pear F1 progeny T003 x ‘Moonglow’ was developed with the purpose of cumulating resistances to fire blight and psylla deriving from Asian and European pear cultivars. Single nucleotide polymorphism (SNP) and simple sequence repeat (SSR)-based genetic maps were built for T003 and ‘Moonglow’. Quantitative Trait Loci (QTLs) were detected for the resistances, demonstrating their polygenic nature. Marker-assisted selection (MAS) can now be applied for these two traits. Furthermore, the segregating population exhibited genetic incompatibilities, and the genomic regions associated with hybrid necrosis were mapped for the first time in pear. Development of molecular markers linked to the lethal genes should allow breeders to avoid crosses leading to incompatible combinations that could affect the expression of important agronomic traits co-segregating with these genes.

As características genéticas das culturas agrícolas podem ser influenciadas pelo ambiente, interferindo na produtividade. Sendo assim, esta pesquisa visa entender como ocorre a interação entre genótipo e ambiente (IGA) e a interação entre quantitative trait locus (QTL) e ambiente (IQA), a fim de fornecer instrumentos que possam melhorar a produtividade. Em destaque, vemos que o modelo de efeitos principais aditivos e interação multiplicativa (AMMI), que considera como aditivos os efeitos principais de genótipo e ambiente e como multiplicativos os efeitos da interação, é uma importante ferramenta que permite estudar a interação com dados de ensaios multi-ambientais e apresentar boas previsões na detecção de QTL para novos ambientes. Para se levar em considera- ção a heterogeneidade da variância do erro ao longo dos ambientes, uma generalização do modelo AMMI é utilizada, o modelo AMMI ponderado ou W-AMMI. Nesta pesquisa, observando os dados resultantes de um experimento com 9 genótipos de milho conduzidos em 20 ambientes e 4 blocos, foram comparadas as análises da IGA utilizando o modelo AMMI e W-AMMI. Com um segundo conjunto de dados, resultantes do cruzamento das variedades de cevada Harrington e TR306, com 141 genótipos conduzidos em 12 ambientes foram comparados os resultados das análises da IGA e IQA utilizando os modelos AMMI e W-AMMI, sendo que foram propostas ponderações por linha (todos os ambientes para determinado genótipo ficam com o mesmo peso) e coluna (todos os genótipos para determinado ambiente ficam com o mesmo peso). / The genetic characteristics of crops can be influenced by the environment, interfering with productivity. This research intends to understand the genotype-by-environment interaction(IGA) and quantitative trait locus (QTL)-by-environment interection (IQA), in order to provide tools that can improve productivity. That the Additive Main Effects and Multiplicative Interactions (AMMI) model, has been widely used to study and understand these interactions and has shown to provide good interpretations of both IGA and IQA, as well as QTL detection. In order to take into account of the heterogeneity of error variance over the environments, a generalization of the AMMI model is used, the weighted AMMI or W-AMMI. In this study, observing the data resulting from an experiment with 9 maize genotypes conducted in 20 environments and 4 blocks, was used to compare the results between the AMMI and W-AMMI models. A second set of data, resulting a cross between the barley varieties Harrington and TR306, with 141 genotypes conducted in 12 environments, was used to compare the AMMI and two versions of the W-AMMI (equal weighs per row; and equal weights per column) models in terms of IGA, IQA and QTL detection.

La plupart des maladies humaines ont une étiologie complexe avec des facteurs génétiques et environnementaux qui interagissent. Utiliser des phénotypes corrélés peut augmenter la puissance de détection de locus de trait quantitatif. Ce travail propose d’évaluer différentes approches d’analyse bivariée pour des traits corrélés en utilisantl’information apportée par les marqueurs au niveau de la liaison et de l’association. Legain relatif de ces approches est comparé aux analyses univariées. Ce travail a étéappliqué à la variation de la densité osseuse à deux sites squelettiques dans une cohorted’hommes sélectionnés pour des valeurs phénotypiques extrêmes. Nos résultats montrentl’intérêt d’utiliser des approches bivariées en particulier pour l’analyse d’association. Parailleurs, dans le cadre du groupe de travail GAW16, nous avons comparé lesperformances relatives de trois méthodes d’association dans des données familiales. / The majority of complex diseases in humans are likely determined by both genetic andenvironmental factors. Using correlated phenotypes may increase the power to map theunderlying Quantitative Trait Loci (QTLs). This work aims to evaluate and compare theperformance of bivariate methods for detecting QTLs in correlated phenotypes by linkageand association analyses. We applied these methods to data on Bone Mineral Density(BMD) variation, measured at the two skeletal sites, in a sample of males selected forextreme trait values. Our results demonstrate the relative gain, in particular for associationanalysis, of bivariate approaches when compared to univariate analyses. Finally, we studythe performances of association methods to detect QTLs in the GAW16 simulated familydata.