Using rice chromosome 1 pseudomolecule as a reference, a minimal tiling path for the euchromatic arms of sorghum chromosome 3 was constructed, in which 23 contigs contain an estimated 57.56 Mb of DNA. A total of 409 EST-STS markers and 255 genetic markers have been mapped onto the euchromatic arms providing excellent integration of the genetic and physical maps. A total of 21 contigs containing 9 ESTSTS and 35 genetic markers have been constructed across the heterochromatin block of sorghum chromosome 3 which comprise 16.57 Mb of DNA. Macrocolinearity between sorghum chromosome 3 and rice chromosome 1 was examined based on the mapped EST-STS markers. Approximately 85% of the EST-STS markers were colinear between these two homeologous chromosomes. Estimates of recombination were also determined, which indicates the existence of recombination cold and hot spots. Microcolinearity between sorghum chromosome 3 and rice chromosome 1 was examined at two different levels. In one case, overlapping sorghum BAC pools orthologous to a 5.1 Mb region of rice chromosome 1 were constructed and sequence skimmed. Alignment of the sorghum skim sequences to the TIGR rice gene models revealed ~62% colinearity between the two orthologous regions. In addition, colinearity between sorghum chromosome 3 and rice chromosome 5 was detected within this region which is likely due to the segmental homology between rice chromosome 1 and rice chromosome 5. Microcolinearity between sorghum and rice was also examined by comparing 2 fully sequenced sorghum chromosome 3 BAC clones to the orthologous region of rice chromosome 1. In this analysis, ~65% colinearity was detected for sorghum BAC 82G24 and ~59% colinearity was detected for sorghum BAC 181g10. Microcolinearity was largely confined to gene coding regions and sequences of exons displayed the highest percent identities. Small-scale gene rearrangements were also detected. Finally, RT-PCR analysis was carried out between a set of colinear and non-colinear genes from sorghum and rice to determine whether the loss of colinearity between orthologous genes resulted in a change in transcriptional regulation. No direct link between loss of colinearity and expression pattern was detected in these experiments.

The reliability of genome analysis and proficiency of genetic manipulation in vivo and in vitro are increased by assignment of linkage groups to specific chromosomes, placement of centromeres, orientation with respect to telomeres, and linear alignment with respect to chromosomal features and dimensions. I undertook five studies aimed at integrating sorghum genomics and cytogenetics at several levels. The results help establish an entirely new "cyto-genomics" resource, impacts of which are likely to be broad. In the first study, I developed a FISH-based karyotyping system for Sorghum bicolor Moench. I used integrated structural genomic resources, including linkage maps and large-insert clonal libraries of sorghum genomic DNA to develop a 17-locus probe cocktail for simultaneous fluorescent in situ hybridization (FISH). This probe enabled facile identification of all chromosome pairs in mitotic chromosome spreads. Perhaps just as important, I established time-efficient means to select sorghum BAC clones for multi-probe FISH. Thus, an integrated cyto-genomics system for sorghum can be constructed without need of chromosome flow sorting or microdissection, both of which are difficult and costly. In the second study, hybridization of DNA clones from 37 different genomic regions enabled the assignment of linkage groups and orientation of linkage maps to chromosomes. Comparisons between genetic and physical distances throughout the genome enabled a new nomenclature for linkage group designation in sorghum. The results provide an integrated nomenclature system of Sorghum bicolor chromosomes and linkage groups. In the third study, I created high-resolution maps by FISH to pachytene bivalents for two linkage groups (B and H), and defined relationships between pericentromeric heterochromatin, centromeres, mapped markers and recombination rates. These relationships will help guide the development and use of sorghum genomics. In the fifth study, I used FISH in two ongoing gene-targeted efforts. For the maturity gene ma5 and fertility restoration gene rfl, I estimated physical lengths between currently available flanking molecular markers. This enables estimation of recombination densities in these regions and assessment of the applicability of map-based and -assisted cloning.

The reliability of genome analysis and proficiency of genetic manipulation in vivo and in vitro are increased by assignment of linkage groups to specific chromosomes, placement of centromeres, orientation with respect to telomeres, and linear alignment with respect to chromosomal features and dimensions. I undertook five studies aimed at integrating sorghum genomics and cytogenetics at several levels. The results help establish an entirely new "cyto-genomics" resource, impacts of which are likely to be broad. In the first study, I developed a FISH-based karyotyping system for Sorghum bicolor Moench. I used integrated structural genomic resources, including linkage maps and large-insert clonal libraries of sorghum genomic DNA to develop a 17-locus probe cocktail for simultaneous fluorescent in situ hybridization (FISH). This probe enabled facile identification of all chromosome pairs in mitotic chromosome spreads. Perhaps just as important, I established time-efficient means to select sorghum BAC clones for multi-probe FISH. Thus, an integrated cyto-genomics system for sorghum can be constructed without need of chromosome flow sorting or microdissection, both of which are difficult and costly. In the second study, hybridization of DNA clones from 37 different genomic regions enabled the assignment of linkage groups and orientation of linkage maps to chromosomes. Comparisons between genetic and physical distances throughout the genome enabled a new nomenclature for linkage group designation in sorghum. The results provide an integrated nomenclature system of Sorghum bicolor chromosomes and linkage groups. In the third study, I created high-resolution maps by FISH to pachytene bivalents for two linkage groups (B and H), and defined relationships between pericentromeric heterochromatin, centromeres, mapped markers and recombination rates. These relationships will help guide the development and use of sorghum genomics. In the fifth study, I used FISH in two ongoing gene-targeted efforts. For the maturity gene ma5 and fertility restoration gene rfl, I estimated physical lengths between currently available flanking molecular markers. This enables estimation of recombination densities in these regions and assessment of the applicability of map-based and -assisted cloning.

Centromeres are responsible for sister-chromatid cohesion, kinetochore formation, and accurate transmission of chromosomes. Rice provides an excellent model for organizational and functional studies of centromeres since several of its chromosomes contain limited amounts of satellite and other repetitive sequences in their centromeres. To facilitate molecular characterization of the centromeres, we screened several BIBAC and BAC libraries of japonica and indica rice, using several centromere-specific repeat elements as probes. The positive clones were identified, fingerprinted and integrated into our whole genome physical map databases of the two rice subspecies. BAC/BIBACbased physical maps were constructed for the centromeric regions of the subspecies. To determine whether the genomic organization of the centromeres has changed since the cultivated rice split from its progenitor and to identify the sequences potentially playing an important role in centromere functions, we constructed a large-insert BIBAC library for the wild progenitor of Asian cultivated rice, O. rufipogon. The library contains 24,192 clones, has an average insert size of 163 kb, and covers 5 x haploid genome of wild rice. We screened the wild rice library with two centromere 8-specific overgo probes designed from the sequences flanking centromere 8 of japonica rice. A BIBACbased map was constructed for wild rice centromere 8. Two of the clones, B43P04 and B15E04, were found to span the entire region of the wild rice centromere and thus selected for sequencing the centromere. By sequencing the B43P09 clone, a 95% genomic sequence of the long arm side of wild rice centromere 8 was obtained. Comparative analysis revealed that the centromeric regions of wild rice have a similar gene content to japonica rice, but the centromeric regions of japonica rice have undergone chromosomal rearrangements at both large scale and nucleotide levels. In addition, although the 155-bp satellite repeats showed dramatic changes at the middle region, they are conserved at the 5' and 3' ends of satellite monomers, suggesting that those regions might have important functional roles for centromeres. These results provide not only new insights into genomic organization and evolution, but also a platform for functional analysis of plant centromeres.

In order to combine previous observations from different sources on sex determination, and to identify sex chromosomes including the major sex determination locus in Nile tilapia, physical and genetic maps based on sex-linked markers and genes (such as sex-linked AFLPs, microsatellites, ovarian aromatase and DMO genes) were integrated and anchored. An accurate physical map using FISH techniques on mitotic cells was developed based on a previous map and 23 tilapia BAC clones previously assigned to linkage groups (LGs) 1, 3, 6, 7, 10 and 12; and on meiotic cells, 2 BAC clones containing the SLAM OniY227 and the dmrt4 gene were mapped. The six linkage groups were then assigned to different chromosomes, but surprisingly, the putative sex LG1 was located to a small submetacentric chromosome and not to the larger subtelocentric chromosome 1, where LG3 was assigned instead. The other LGs were assigned to different chromosomes and oriented with respect to the centromeres. A detailed comparison of the physical distribution of markers on chromosome 1 with respect to LG3 revealed a suppression in recombination in the subtelomeric region of the q arm between the marker GM354 (0 cM) and clcn5 (29 cM) and an abrupt increment of recombination between clcn5 (29 cM) and GM128 (77 cM) close to the centromere (Flpter=0.2). The unpairing region (20% of the total length) observed on the larger bivalents of XY fish during early pachytene in meiotic cells has been confirmed by DAPI staining and FISH to be at the terminal part of the q arm, opposite to the centromere. Comparison with six other tilapia species (2n=44) revealed a well conserved karyological distribution of the suspected LGs associated with sex determination (1 and 3). Besides, in O. karongae (2n=38) it was shown by SATA and UNH995/UNH104 marker hybridisation that LG1 has been re-arranged into the subtelomeric chromosome 2 as a result of a telomere-telomere fusion. A pool of 15 tilapia BAC clones previously localised on chromosome 1 and containing sex-linked AFLPs, dmrt1, dmrt4 and several SINEs were screened for new microsatellites; BACs were digested with SAU3AI and TC, GT, ATCT and CTGT probes radio-labelled with 32P. The high abundance of repetitive sequences in the BACs used led to only one useful polymorphic and co-dominant marker being obtained, associated to a BAC clone containing a copy of the dmrt1 gene on chromosome 1 (Flpter=0.85). Four linkage maps were constructed from an XY male, XY neofemale, XX neomale and XX female, mapping 4 and 8 markers on LG1 and LG3 (including the dmrt1 associated microsatellite) respectively. A specific sex-determination locus was identified on LG1 clearly linked with UNH995. However there appeared to be different allelic strengths for this sex determination locus, as shown by different sex ratios associated with different UNH995 genotypes. Additionally, one of the two XX fish mapped, showed the location of the recessive black blotching trait on LG3 (chromosome 1) between the markers GM128 and GM526, close to the centromere (Flpter=0.14). The results presented suggest a nascent Y chromosome in early stage of differentiation in Nile tilapia and with a functional master gene on LG1 close to the marker UNH995 (Flpter=0.67) located on the q arm of a small submetacentric chromosome. The potential influences of the autosomal LG3 (chromosome 1) in sex differentiation are also discussed.

The Medicago truncatula NIP gene is located on M. truncatula Linkage Group 1. Informative recombinants showed crossovers that localize the NIP gene between markers 146O17 and 23C16D. Marker 164N9 co-segregates with the NIP gene, and the location of marker 164N9 is between markers 146O17 and 23C16D. Based upon data from the Medicago genome sequencing project, a subset of the model legume Medicago truncatula bacterial artificial chromosomes (BACs) were used to create a physical map on the DNA in this genetic internal. BACs near the potential NIP gene location near marker 164N9 were identified, and used in experiments to predict the physical map by a BAC-by-BAC strategy. Using marker 164N9 as a center point, and chromosome walking outward, the physical map toward markers 146O17 and 23C16D was built. The chromosome walk consisted of a virtual walk, made with existing sequence of BACs from the Medicago genome project, hybridizations to filters containing BAC DNA, and PCR reactions to confirm that predicted overlapping BACs contained DNA that yielded similar PCR products. In addition, the primers which are made for physical mapping via PCR could be good genetic markers helpful in discovering the location of the NIP gene. As a result of efforts repotted here, gap in physical map between marker 164N9 and 146O17 was closed.

Amaranthus hypochondriacus is an emerging pseudo-cereal native to the New World which has garnered increased attention in recent years due to its nutritional quality, in particular its seed protein, and more specifically its high levels of the essential amino acid lysine. It belongs to the Amaranthaceae family, is an ancient paleotetraploid that shows amphidiploid inheritance (2n=32), and has an estimated genome size of 466 Mb. Here we present a high-quality draft genome sequence of the grain amaranth A. hypochondriacus. The genome assembly consisted of 377 Mb in 3,518 scaffolds with an N50 of 371 kb. Repetitive element analysis predicted that 48% of the genome is comprised of repeat sequences, of which Copia-like elements were the most common classified retrotransposon. A transcriptome, consisting of 66,370 contigs, was assembled from eight different tissue and abiotic stress libraries. Annotation of the genome identified 23,059 genes that were supported by our de novo transcriptome assembly, the RefBeet 1.1 gene index and the Uniprot_sprot database. To describe the genetic diversity within the grain amaranths (A. hypochondriacus, A. caudatus, and A. cruentus) and their putative progenitor (A. hybridus) we re-sequenced seven accessions in the genus Amaranthus (four A. hypochondriacus, and one of each A. caudatus, A. cruentus, and A. hybridus), which identified 7,184,636 and 1,760,433 interspecific and intraspecific single nucleotide polymorphisms, respectively. A phylogeny analysis of the re-sequenced accessions substantiated the classification of A. hybridus as the progenitor species of the grain amaranths. Lastly, we generated a physical map for A. hypochondriacus using the BioNano optical mapping platform. The physical map spanned 340 Mb and a hybrid assembly using the BioNano optical genome maps nearly doubled the N50 of the assembly to 697 kb. Moreover, we analyzed synteny between amaranth and Beta vulgaris (sugar beet) and estimated, using Ks analysis, the age of the most recent polyploidization event in amaranth.

The Culicinae is the most extensive subfamily among the Culicidae family of mosquitoes. Two genera, Culex and Aedes, from this subfamily have world-wide distribution and are responsible for transmitting of several deadly diseases including Zika, West Nile fevers, chikungunya, dengue, and Rift Valley fevers. Developing high-quality genome assembly for mosquitoes, studying their population structure, and evolution can help to facilitate the development of new strategies for vector control. Studies on Aedes albopitcus as well as on species from the Culex pipiens complex, which are widely spread in the United States, provide excellent models on these topics. Ae. albopictus is one of the most dangerous invasive mosquito species in the world that transmits more than 20 arboviruses. This species has highly repetitive genome that is the largest among mosquito genomes sequenced so far. Thus, sequencing and assembling of such genome is extremally challenging. As a result, the lack of high-quality Ae. albopictus genome assembly has delayed the progress in understanding its biology. To produce a high-quality genome assembly, it was important to anchor genomic scaffolds to the cytogenetic map creating a physical map of the genome assembly. We first developed a new gene-based approach for the physical mapping of repeat-rich mosquito genomes. The approach utilized PCR amplification of the DNA probes based on complementary DNA (cDNA) that does not include repetitive DNA sequences. This method was then used for the development of a physical map for Ae. albopictus based on the in situ hybridization of fifty cDNA fragments or gene exons from twenty-four scaffolds to the mitotic chromosomes from imaginal discs. This study resulted in the construction of a first physical map of the Ae. albopictus genome as well as mapping viral integration and polyphenol oxidase genes. Moreover, comparing our present Ae. albopictus physical map to the current Ae. aegypti assembly indicated the presence of multiple chromosomal inversions between them. To better understand population structure and chromosome evolution in Culicinae mosquitoes, especially in the Culex pipiens complex, we studied genomic and chromosomal differentiation between two subspecies Cx. pipiens pipiens and Cx. pipiens molestus. For the species responsible for the spread of human diseases, understanding the population dynamics and processes of taxa diversification is important for an effective mosquito control . Two vectors of West Nile virus, Cx. p. pipiens and Cx. p. molestus, exhibit epidemiologically important behavioral and physiological differences, but the whole-genome divergence between them was unexplored. The first goal of this study was to better understand the level of genomic differentiation and population structures of Cx. p. pipiens and Cx. p. molestus from different continents. We sequenced and compared whole genomes of 40 individual mosquitoes from two locations in Eurasia and two in North America. Principal Component, ADMIXTURE, and neighbor joining analyses of the nuclear genomes identified two major intercontinental, monophyletic clusters of Cx. p. pipiens and Cx. p. molestus. The level of genomic differentiation between the subspecies was uniform along chromosomes. The ADMIXTURE analysis determined signatures of admixture in Cx. p. pipens populations, but not in Cx. p. molestus populations. Thus, our study identified that Cx. p. molestus and Cx. p. pipiens represent different evolutionary units with monophyletic origin that have undergone incipient ecological speciation. The second goal was to study differences at the chromosome level between these two organisms. We first measured whole chromosome and chromosome arm length differences between Cx. p. molestus and Cx. p. pipiens as a basic cytogenetic approach. In addition, we used the novel Hi-C approach to detect chromosomal rearrangements between them since Hi-C was successful in detecting a known inversion in Cx. quinquefasciatus. Cx. p. molestus and Cx. p. pipiens embryos were used to perform the Hi-C technique. Analysis of the Hi-C data showed the presence of two different inversions in Cx. p. pipiens and Cx. p. molestus heatmap, which could explain their different physiology and adaptation in nature. Developing modern genomic and cytogenetic tools is important to enhance the quality of genome assemblies, improve gene annotation, and provide a better framework for comparative and population genomics of mosquitoes; also it is the foundation for the development of novel genome-based approaches for vector control. / Doctor of Philosophy / Mosquitoes are medically important insects because they vector a range of diseases that infect humans. The subfamily Culicinae is responsible for transmitting such diseases as Zika, dengue, and West Nile fevers, which have triggered fatal infections and epidemics in multiple parts of the world. Since 2010-2016, studies have reported exceeding levels of insecticide resistance that slows the disease elimination process. Novel transgenic techniques have a tremendous potential for more efficiently minimizing mosquito-borne diseases and transmission. Availability of high-quality genome assemblies for mosquitoes may help to better understand their population structure and to develop effective and safe vector-control approaches that we urgently need. For the development of high-quality genome assemblies, we need to construct a physical genome map, that shows the physical locations of genes or other DNA sequences of interest along the chromosomes. For this reason, we developed a new gene-based approach for the physical mapping of the mosquito genomes. This method was then used for the development of a physical map for Ae. albopictus. This study resulted in the generation of the first physical map of the Ae. albopictus genome. To understand population structure in Culicinae mosquitoes, we used mosquitoes from the Culex pipiens complex. Species in this complex transmit different arthropod-borne viruses or arboviruses. Notable is the West Nile Virus, which has triggered fatal infections and epidemics in Eastern and Central Europe, North America and is also known in Asia, Australia, Africa, and the Caribbean. We specifically focused on two subspecies in this complex, Cx. pipiens pipiens and Cx. pipiens molestus that are morphologically identical, but are different physiologically and behaviorally. Although they are spread globally in temperate regions, their population structure and taxonomic status remains unclear. The first goal of this study was to better understand the level of genomic differentiation of Cx. p. pipiens and Cx. p. molestus from different continents. We sequenced and compared the whole genomes of 40 individual mosquitoes from two locations in Eurasia and two in North America. Our study identified that Cx. p. molestus and Cx. p. pipiens represent different evolutionary units that are currently undergoing ecological speciation. The second goal was to study differences at the chromosome level between them. Using the Hi-C approach we detected presence of two different inversions in Cx. p. pipiens and Cx. p. molestus, which could potentially explain their different physiology and adaptation.

La qualité du fruit est un critère incontournable de sélection chez les Rosacées fruitières, et l’acidité constitue une composante majeure de la qualité organoleptique. Toutefois, les mécanismes physiologiques et moléculaires contrôlant l'acidité des fruits restent mal connus. Chez le pêcher, le caractère non acide du fruit est contrôlé par le locus D. Une descendance F2 de 208 individus issus d'un croisement entre une variété de pêche non acide ‘Ferjalou Jalousia®’ et une variété de nectarine normalement acide, ‘Fantasia’ (JxF) a été analysée pour différents caractères agronomiques dont l’acidité du fruit. Cette descendance a servi à la réalisation d’une carte génétique et ainsi à la localisation du locus D sur le groupe de liaison 5 (GL5). Ce locus co-localise avec des QTL à effet majeur impliqués dans l’acidité titrable, le pH, la teneur en acides organiques ainsi que des QTL à effet plus faible pour la teneur en sucres solubles. De nombreux gènes candidats impliqués dans la synthèse des acides organiques, la dégradation et le stockage vacuolaire avaient été précédemment étudiés. Cependant, aucun gène candidat n’a encore été cartographié dans la région du locus D, excluant ainsi leur rôle direct dans le contrôle de l’acidité du fruit. Ceci s’explique par la complexité des voies métaboliques des acides organiques et par l’implication de transporteurs, de canaux et de pompes à protons qui rendent l'identification du ou des gène(s) associés au locus D plus complexe par une approche gène candidat. Par conséquent, une approche de clonage positionnel a été menée dans la présente étude afin d’identifier le ou les gène(s) intervenant dans le contrôle de l’acidité du fruit chez le pêcher dans le but de comprendre les mécanismes moléculaires et physiologiques sous-jacents. La recherche de marqueurs liés au locus D a été réalisée par BSA-AFLP. Trente quatre marqueurs AFLP ont été cartographiés sur le GL5 et les six marqueurs les plus proches ont été convertis en marqueurs SCAR codominants. Une carte génétique fine de la région contenant le locus D a ensuite été réalisée à partir d’une descendance F2 élargie à 1 718 individus et à l’aide des six marqueurs SCAR et de trois marqueurs microsatellites précédemment cartographiés dans cette région. L’ensemble des données de génotypage et de phénotypage des individus ayant subi un événement de recombinaison dans la région du locus d’intérêt, a permis la localisation précise du locus D dans un intervalle de 0,4 cM. En parallèle, une banque BAC d’une couverture estimée à 15 fois la taille du génome haploïde du pêcher a été réalisée et son criblage a permis d’évaluer le rapport distance physique/distance génétique dans cette région à 250 kb/cM. Après deux étapes de marche sur le chromosome, une carte physique de la région a été construite en intégrant 16 marqueurs issus du séquençage d’extrémités de BAC. Un clone BAC de 98 kb contenant l’allèle D et un autre de 78 kb contenant l’allèle d ont été séquencés. L’annotation des séquences des deux allèles a mis en évidence onze gènes candidats. De nouveaux marqueurs développés à partir des séquences de ces deux BAC ont ensuite permis de préciser la localisation du locus d’intérêt dans un intervalle de 16 kb. Dans cette région deux gènes ont été identifiés : un gène de résistance et un gène codant pour un transporteur. Une approche transcriptionnelle a été initiée en complément du clonage positionnel afin de fournir un premier élément pouvant confirmer l’implication d’un ou plusieurs gène(s) candidat(s) positionnel(s) dans l’acidité du fruit chez le pêcher. / Acidity is an essential component of the organoleptic quality of fleshy fruits. However, the physiological and molecular mechanisms that control fruit acidity remain unclear. In peach low-acidity is determined at the D locus by the dominant allele. A peach progeny of 208 F2 individuals obtained from a cross between ‘Ferjalou Jalousia®’ (a low-acid peach) and ‘Fantasia’ (a normally acid nectarine) varieties (JxF) was analyzed for several agronomical traits. This peach F2 progeny segregating for several mendelian traits, was analyzed for fruit quality traits including fruit acidity and used for the construction of a genetic linkage map. The D locus was mapped to the proximal end of linkage group 5 (LG5) and co-localized with major QTLs involved in the control of fruit pH, titratable acidity and organic acid concentration and minor QTLs for sugar concentration. Several candidate genes involved in organic acids synthesis, degradation or vacuolar storage have previously been studied. However, none of these candidate genes were located on in the region of the D locus, excluding their direct role in the control of fruit acidity by the D locus. The complexity of organic acids metabolic pathways as well as the involvement of transporters and channels and related proton pumps has hampered, so far, the identification of the gene(s) associated to the D locus using a candidate gene approach. Thus, in order to investigate the molecular and physiological bases of fruit acidity in peach, a positional cloning strategy of the D locus was undertaken for the isolation of the gene(s) underlying this trait. Using a BSA-AFLP method, 34 AFLP markers were mapped to the LG5, and the six nearest markers were transformed into codominant SCAR markers. These SCAR markers and three previously mapped SSR markers were used to genotype an F2 segregating progeny extended to 1,718 F2 individuals. A high-resolution map of the D locus was realized after genotyping and phenotyping recombinant individuals. Using these recombinant plants we delimited the D locus to a genetic interval of 0.4 cM. We also constructed a peach BAC library with a covering estimated at 15 x the peach haploid genome. The screening of the BAC library with tightly linked markers indicated that 1 cM corresponds to 250 kb at the vicinity of the D locus and allowed the construction of the physical map in two walks integrating 16 markers obtained from the BACends sequences. Two BAC clones harbouring the D locus were identified and sequenced; one BAC clone of 98 kb containing the D dominant allele and another one of 78 kb containing the d recessive allele. Eleven predicted genes were found in the sequenced region. A new set of markers was developed which allowed the localization of the D locus in a 16 kb interval. In this region, two genes were identified: a resistance gene and a gene encoding for a transporter. A transcriptional approach was initiated in addition to the positional cloning strategy to provide a first element which could confirm the involvement of one or more identified positional candidate gene(s) in the control of peach fruit acidity.

Des épisodes climatiques de sècheresse et/ou de hautes températures peuvent engendrer de fortes pertes de rendement pour les cultures de céréales au champ. Un QTL associé au rendement et à ses composantes a été détecté dans quatre populations de blé (Triticum aestivum L.) sur le bras long du chromosome 3B « qYDH.3BL ». Deux populations d’haploïdes doublés (RAC875/Kukri et Excalibur/Kukri) et deux populations de lignées recombinantes (RAC875/Kukri et Gladius/Drysdale) ont été utilisées pour cartographier finement le QTL, au même titre que l’identification de gènes candidats. Ces quatre populations ont été testées sous des conditions environnementales variées, incluant des périodes de sécheresse et/ou hautes températures en Australie et au Mexique. Des modèles statistiques mixtes et linéaires décomposant les variations génétiques et non-génétiques ont été utilisés pour la détection de QTL en considérant dans un premier temps chaque environnement unique, puis en considérant les environnements multiples dans une analyse commune. Les allèles de RAC875, Drysdale et Excalibur à ce locus ont montré une hausse du rendement de 5 à 12.5 % comparées à celles de Gladius ou Kukri. Un total de trente-sept marqueurs moléculaires a été cartographié dans la région du QTL. Les marqueurs moléculaires ont été sélectionnés (i) par comparaison avec une carte génétique publiée du chromosome 3B, ou (ii) en désignant de nouveaux marqueurs moléculaires sur les séquences de BAC-end, de contig ou de gènes provenant du projet de séquençage du chromosome 3B (3BSEQ, http://urgi.versailles.inra.fr/, cv. Chinese Spring). Ceci a permis la construction d’une carte génétique consensus du locus qYDH.3BL . A ce jour, aucun QTL associé au rendement ou ses composantes en condition de sécheresse et/ou de hautes températures n’a encore été cloné positionellement chez le blé tendre. Les marqueurs moléculaires de la région d’intérêt ont été utilisés pour cartographier physiquement des contigs, soit par PCR, soit par comparaison de séquences in silico. La région du QTL inclus un total de huit contigs physiques comprenant 85 gènes annotés. L’utilisation de base de données de transcris biologiques publiques ou internes ont été utilisées pour détecter la présence de ces gènes, réduisant la liste à soixante-cinq gènes. Sur les contigs ayant une confiance élevée, aucun des vingt gènes n’a été exprimé différentiellement entre RAC875 et Kukri. Cependant, un gène présentant du polymorphisme dans sa séquence ainsi qu’une délétion/insertion d’un segment portant 12 gènes ont été découvert permettant ainsi de continuer à affiner la liste de gènes candidats. Les trois lignées parentales (RAC875, Drysdale et Excalibur) qui ont l’allèle liée au haut rendement ont le même haplotype pour ce gène, et la même délétion/insertion en opposition au deux autres lignées parentales Gladius et Kukri. Ainsi, dans ce travail de thèse nous avons pu confirmer la présence d’un QTL répondant aux stresses environnementaux sur le chromosome 3BL dans différentes populations et différents environnements, identifier des gènes candidats sous le QTL, et proposer une liste restreinte pour de futures analyses sur la base de données d’expression et de polymorphismes entre les parents des populations de cartographie. / Drought and heat can occur during the growth cycle of crops and severely reduce yield. A QTL associated with yield and yield-related component was found in four wheat populations (Triticum aestivum L.) on the long arm of chromosome 3B “qYDH.3BL”. The four populations were grown under various climatic conditions including drought, heat and combinations of both in a number of different areas (Australia and Mexico). Linear mixed models that partition and account for genetic and non-genetic or extraneous variation were used to detect loci in single-environment and/or multi-environment QTL analysis using ASReml-R. The alleles carried by RAC875, Excalibur or Drysdale improved grain yield by between 5% and 12.5%. Two doubled haploid populations (RAC875/Kukri and Excalibur/Kukri) and two recombinant inbred line populations (RAC875/Kukri and Gladius/Drysdale) were used to fine map qYDH.3BL and identify candidate gene(s). A total of thirty-seven molecular markers were mapped on one or both genetic maps of chromosome 3B enabling development of a consensus genetic map of the qYDH.3BL region. The markers were selected based on comparisons with a published “neighbour map” of chromosome 3B or designed using either BAC-end, contig or gene sequences from the chromosome 3B sequencing project; 3BSEQ http://urgi.versailles.inra.fr/ (cv. Chinese Spring). A positional cloning approach was used to identify candidate genes for qYDH.3BL. Molecular markers from the targeted region were assigned to physical contigs by screening the chromosome 3B BAC library experimentally using PCR or in silico by sequence comparison. A total of eight physical contigs containing 85 genes, were anchored to the qYDH.3BL region. Public and in-house resources of wheat transcript sequences were used to restrict the gene list to 65 expressed genes. Based on comparison of the 65 gene sequences to gene probes in a drought transcriptomic database, three genes were found to be differentially expressed between RAC875 and Kukri under drought conditions. Short genomic sequence reads (10× coverage) from each of the five parental lines (RAC875, Kukri, Excalibur, Gladius and Drysdale) were mapped against the 65 genes for polymorphism discovery. One gene exhibited sequence polymorphism between the drought tolerant parents (RAC875, Excalibur and Drysdale) and the drought-sensitive parents (Gladius and Kukri). In addition, presence/absence polymorphisms were consistently detected throughout a region containing 12 genes, indicating that the drought tolerant parents may have a deletion (or alien introgression) in this region. Thus, in this work, we confirmed the genetic effect of qYDH.3BL in multiple environments and multiple populations, saturated the target region with new molecular markers and defined a preliminary list of genes located in the qYDH.3BL region and selected candidate genes for further investigations.