Tambaqui (Colossoma macropomum) is the largest native Characiform species from the Amazon and Orinoco river basins of South America. Tambaqui farming is growing rapidly in Brazil, its production reached 139.209 tons in 2014, what corresponds to 57.7% of increase compared with 2013. However, few genetic studies of tambaqui are currently available. The tambaqui genetic studies for cultured and wild populations need a holistic approach for a rational action facing ecological and market challenges in aquaculture. Approaches based on genetic studies have provided important tools to understand population dynamics, local adaptation, and gene function to improve selection strategies to be applied in breeding programs. The next-generation sequencing (NGS) allowed a great advance in genomic and transcriptomic approaches, especially related to non-model species. The genotype-by-sequencing (GBS) is one of this approaches based on genome complexity reduction using restriction enzymes (REs). This thesis presents the application of these approaches to provide advances in the genetic background for tambaqui studies. The GBS approach provided a high-density SNPs panel that allowed us to develop the first linkage map, and association studies with environmental variables, local adaptation, and lack of intermuscular bones, both using tambaqui as a model. This work can give us many theoretical references to be applied in genetic breeding programs for tambaqui, allowing a better understanding of genetic processes related to traits of interest in aquaculture. / O tambaqui (Colossoma macropomum) é a maior espécie nativa de Characiforme da América do Sul e é encontrado nas bacias do rio Amazonas e Orinoco. O cultivo do tambaqui está crescendo rapidamente no Brasil, sua produção atingiu 139.209 toneladas em 2014, o que corresponde a 57,7% de aumento em relação a 2013. No entanto, poucos estudos genéticos realizados com o tambaqui estão disponíveis atualmente. Estudos genéticos em tambaqui, tanto em populações cultivadas quanto em populações selvagens, necessitam de uma abordagem holística para uma ação racional frente aos desafios ecológicos e mercadológicos na aquicultura. Abordagens baseadas em estudos genéticos têm fornecido ferramentas importantes para se entender a dinâmica populacional, adaptação local e função gênica visando melhorar as estratégias de seleção a serem aplicadas em programas de melhoramento genético. O sequenciamento de nova geração (NGS) permitiu um grande avanço nas abordagens genômicas e transcriptômicas, especialmente relacionadas a espécies não-modelo. A genotipagem por sequenciamento (GBS) é uma dessas abordagens que utilizam enzimas de restrição (REs) para reduzir a complexidade do genoma. Esta tese apresenta a aplicação desta abordagem objetivando proporcionar avanços significativos nos estudos genéticos de base para tambaqui. A técnica de GBS forneceu um painel de SNPs de alta densidade que nos permitiu desenvolver o primeiro mapa de ligação e estudos de associação com variáveis ambientais, adaptação local e ausência de ossos intermusculares no tambaqui. Este trabalho pode nos dar muitas referências teóricas a serem aplicadas em programas de melhoramento genético do tambaqui, permitindo uma melhor compreensão dos processos genéticos relacionados a traços de interesse na aquicultura.

The size and complexity of the oat genomes (Avena L., x = 7) have made genetic studies, including the discovery of molecular markers, difficult. Recent attention to these species has resulted in the development of many DArT -based markers in the tetraploid A. magna Murphy et Terrill (2n = 28, CCDD genomes), along with numerous RFLP's, SSR's, DArT's, and EST-based SNPs in hexaploid A. sativa L. (2n = 42, AACCDD). Here we report the first SNP markers for tetraploid oat based on genome reduction and high-throughput pyrosequencing in two inbred lines of A. magna: A-169 (wild) and Ba 13-13 (domesticated). Initially, the genomes were reduced using restriction digests with EcoRI and BfaI and sequenced to produce 706,426 reads for both genotypes that were subsequently assembled into 57,048 contigs with an average read length of 345 bp. Comparisons of the contigs between the two lines resulted in the detection of 31,304 in silico SNPs. High Resolution Melt (HRM) and KASPar assays were used to validate 1,108 of these in silico SNPs across a panel of diploid, tetraploid, and hexaploid oats. Of the assays, 119 were validated using HRM and 384 using KASPar genotyping in the Fluidigm EP1 system. Both sets of assays were then mapped on a population of 117 F2:8 recombinant inbred lines (RILs) developed from the A-169 x Ba 13-13 cross. A map of the A. magna genome was then constructed. The markers and map provide a new set of genomic tools for tetraploid and hexaploid oat breeding and allow for tracking of genes controlling traits of economic importance and other interesting genes through the evolution of Avena.

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

Thesis (PhD(Agric))--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Grapevines are one of the most important and diverse crops in the world, but tend to be susceptible for numerous pests and diseases. The dagger nematode, Xiphinema index (X. index) is a well-known soil-borne pest of grapevine and vector of grapevine fanleaf virus. Several Vitis species showed resistance to this pest. Breeding efforts have been underway for several decades to create resistant rootstocks. However, conventional breeding efforts are time consuming due to grapevines being a perennial crop, its heterozygosity, as well as its long growth cycle. Breeding new grapevine varieties are also expensive and work intensive. The development of marker-assisted selection introduced a way to overcome some of the abovementioned problems. The aim of this study was to broaden the genetic evaluation and breeding efforts for improved X. index resistance in grapevine rootstocks. In 2007 several crosses were made in the University of California, Davis vineyards. The background for all these crosses consisted of V. arizonica. These V. arizonica plants are part of a collection obtained by H.P. Olmo during the 1960’s. In recent studies it was established that X. index resistance is controlled by a single dominant gene. The 0701 (R8916-07 (Wichita Refuge x b40-14) x R8916-32), 0704 (161-49C x b40-14) and 0705 (161-49C x R8916-22) populations were created to confirm the homozygous nature of b40-14, a V. arizonica accession. In addition, several V. arizonica species were screened to confirm their resistance or susceptibility towards X. index feeding. The 0705 population was also used to create a genetic map for X. index resistance. In this study a new and improved screening method was developed to inoculate vines under greenhouse conditions. This screening method proved to be quicker and less damaging on the nematodes than traditional systems. Control varieties were used and O39-16, a commercial rootstock showed no damage, even with high nematode pressure, whereas V. rupestris Saint George had severe root damage and decline after eight weeks of exposure. A range of V. arizonica accessions was tested for their resistance to X. index feeding. Of the 18 genotypes tested, half showed resistance and the rest were susceptible. It is possible that these genotypes are not pure V. arizonica genotypes. Genotypes with V. arizonica in the background were also tested. Wichita Refuge was used as a susceptible female parent and the progeny were expected to be heterozygous resistant. Some of the progeny allowed low levels of feeding damage, which may have been the result of the more effective inoculation method described above. The 0701 population confirmed the hypothesized model of 3:1 (Resistance (R):Susceptible (S)) segregation although 13 of the genotypes showed significantly higher gall numbers than the susceptible female parent. The possibility of transgressive segregation exists, but needs to be confirmed. All progeny from the 0704 population should be heterozygous resistant, but a 1:1 (R:S) segregation pattern was observed. The 0705 population was created as a mapping population to study X. index resistance. This population was also tested in the greenhouse for its X. index resistance and was expected to segregate 1:1 (R:S). The X2 analysis did not fully support this model. A genetic map covering all 19 linkage groups, and positioning 175 polymorphic SSR markers was created for the 164 progeny in the 0705 population. MapQTL analysis revealed a major QTL on linkage group 9 and two minor QTL’s on groups 13 and 19. The major QTL placed between markers VMC1c10 and CTG1032918 with a LOD score of 33.4 explaining 70.5% of the phenotypic variance for X. index. This QTL is the second major QTL discovered for X. index resistance. With the discovery of a second major QTL, the two types of resistance can be pyramided. Work is underway to saturate the area around the major QTL on linkage group 9 and to move towards physical mapping of X. index resistance. The b40-14 V. arizonica accession is also known for its resistance to Pierce’s disease and the possibility of simultaneous expression of two types of resistance is created. The 0705 population can also be used to evaluate phenotypical characteristics in the field to determine if useful rootstocks can be selected. Taken together, the results obtained in this study provide improved methods and highly characterized plant populations to support the efforts in obtaining improved X. index resistance in grapevine rootstocks. / AFRIKAANSE OPSOMMING: Wingerde is van die belangrikste en mees diverse gewasse op aarde, maar hulle neig om vir ‘n verskeidenheid plae en siektes vatbaar te wees. Die dolk-aalwurm, Xiphinema index (X. index), is ‘n bekende grondgedraagde plaag van wingerd en ‘n vektor vir wingerd- netelblaarvirus. Verskeie Vitis-spesies toon weerstand teen hierdie plaag. Daar word reeds vir dekades pogings aangewend om weerstandbiedende onderstokke te kweek. Konvensionele kweekpogings is egter tydrowend omdat wingerd ‘n meerjarige gewas is, op grond van die heterosigositeit van die gewas, sowel as die lang groeisiklus. Dit is ook duur en arbeidsintensief om nuwe wingerdvariëteite te kweek. Die ontwikkeling van merkerondersteunde seleksie het dus ‘n metode verskaf om sommige van bogenoemde probleme te oorkom. Die doelwit van hierdie studie was om die genetiese evaluerings- en kweekpogings vir verbeterde X. index-weerstand in wingerd-onderstokke te verbreed. In 2007 is verskeie kruisings in die wingerde by die Universiteit van Kalifornië, Davis gemaak. Die agtergrond vir al hierdie kruisings het bestaan uit V. arizonica. Hierdie V. arizonica-plante vorm deel van ‘n versameling wat in die 1960’s deur H.P. Olmo verkry is. In onlangse studies is daar bepaal dat X. index-weerstand deur ‘n enkele dominante geen beheer word. Die 0701 (R8916-07 (Wichita Refuge x b40-14) x R8916-32), 0704 (161-49C x b40-14) en 0705 (161-49C x R8916-22) bevolkings is geskep om die homosigotiese geaardheid van b40-14, ’n V. arizonicaafstammeling, te bevestig. Daarbenewens is verskeie V. arizonica-spesies gesif om hulle weerstand teen of vatbaarheid vir X. index voeding te bevestig. Die 0705 bevolking is ook gebruik om ‘n genetiese kaart vir X. indexweerstand te skep. In hierdie studie is ‘n nuwe en verbeterde siftingsmetode ontwikkel om wingerdstokke onder glashuistoestande te inokuleer. Daar is gewys dat hierdie siftingsmetode vinniger en minder skadelik vir die aalwurms as tradisionele metodes is. Beheervariëteite is gebruik en O39-16, ‘n kommersiële onderstok, het geen skade getoon nie, selfs met hoë aalwurmdruk, terwyl V. rupestris Saint George ernstige wortelskade en agteruitgang na agt weke se blootstelling getoon het. ‘n Verskeidenheid V. arizonica-afstammelinge is vir hulle weerstand teen X. index-voeding getoets. Van die 18 genotipes wat getoets is, het die helfte weerstand getoon en die res was vatbaar. Dit is moontlik dat hierdie genotipes nie suiwer V. arizonica-genotipes was nie. Genotipes met V. arizonica in hulle agtergrond is ook getoets. Wichita Refuge is as ‘n vatbare vroulike ouer gebruik en die verwagting was dat die nageslag heterosigoties weerstandbiedend sou wees. Sommige van die nageslag het lae vlakke van voedingskade toegelaat, wat moontlik die gevolg was van die meer doeltreffende inokulasiemetode wat hierbo beskryf word. Die 0701 bevolking het die veronderstelde model van 3:1 (Weerstandbiedend (W):Vatbaar (V)) segregasie bevestig, hoewel 13 van die genotipe noemenswaardig hoër galgetalle as die vatbare vroulike ouer getoon het. Die moontlikheid van transgressiewe segregasie bestaan, maar dit moet nog bevestig word. Alle nageslag van die 0704 bevolking behoort heterosigoties weerstandbiedend te wees, maar ‘n 1:1 (W:V) segregasiepatroon is waargeneem. Die 0705 bevolking is as ‘n karteringsbevolking geskep om X. index-weerstand te bestudeer. Hierdie bevolking is ook in die glashuis vir sy X. index-weerstand getoets en daar is verwag dat dit 1:1 (W:V) sou segregeer. Die X2 analise het nie hierdie model ten volle ondersteun nie. ‘n Genetiese padkaart wat al 19 skakelingsgroepe en die posisies van 175 polimorfiese SSR merkers toon, is vir die 164 afstammelinge in die 0705 bevolking geskep. MapQTL analise het ‘n groot kwantitatiewe eienskap lokus (QTL) op skakelingsgroep 9 en twee kleiner QTL’e op groepe 13 en 19 onthul. Die groot QTL is tussen merkers VMC1c10 en CTG1032918 met ‘n LOD telling van 33.4 geplaas en het 70.5% van die fenotipiese variansie van X. index verklaar. Hierdie QTL is die tweede groot QTL wat vir X. index-weerstand ontdek is. Met die ontdekking van ‘n tweede groot QTL, kan die twee soorte weerstand gepiramideer word. Werk word reeds onderneem om die area rondom die groot QTL op skakelingsgroep 9 te versadig en om na die fisiese kartering van X. index-weerstand te beweeg. Die b40-14 V. arizonica-afstammeling is ook bekend vir sy weerstand teen Pierce se siekte en die moontlikheid word geskep vir die gelyktydige uitdrukking van twee soorte weerstand. Die 0705 bevolking kan ook gebruik word om die fenotipiese kenmerke in die veld te evalueer om te bepaal of bruikbare onderstokke geselekteer kan word. In kombinasie behoort die resultate wat in hierdie studie verkry is, verbeterde metodes en hoogs gekarakteriseerde plantbevolkings te lewer wat die pogings sal ondersteun om verbeterde X. index-weerstand in wingerdonderstokke te verkry.

Acca sellowiana, known as feijoa or pineapple guava, is a Myrtaceae fruit tree species native to Uruguay and Brazil. The species stand out for its highly aromatic fruits, with nutraceutical and therapeutic value. Despite its agronomically promising valuable, genetics studies on this species are limited. Linkage genetic maps are valuable tools for genetic and genomic studies, and can be employed in breeding programs to support the development of molecular breeding strategies. The lack of a high number of polymorphic markers is one of the main limitation to development saturated genetic maps. Recently, novel genotyping methods based on next generation sequencing technology allow to detect and genotype thousands of markers in mapping populations. This represents a rapid and cost-effective strategy, remarkably useful for minor species with limited genomic resources. In this study, we constructed a high-density integrated genetic linkage map of A. sellowiana using two populations, H5 (\'TCO × BR\' , n = 160) and H6 (\'TCO × DP\', n = 184), which have the same female parent. Genotyping by sequencing (GBS) approach was used to simultaneously discover and genotype single nucleotide polymorphism (SNP) markers in both populations. Two strategies were carried out to identify SNP markers: a reference pipeline using the reference genome of the closely-related species Eucalyptus grandis, and a de-novo pipeline that do not require a reference genome. After quantitative genotype calling, 5,350 and 4,227 high quality SNP markers were selected for mapping in H5 and H6 populations, respectively. The two resulting maps of populations H5 and H6 comprised 1,236 and 1,302 markers distributed over the expected 11 linkage groups. The H5 and H6 maps spanned a map length of 1,593 cM and 1,572 cM, with an average inter-marker distance of 1:29 cM and 1:21 cM, respectively. A high degree of collinearity was observed between the two maps. In addition, a large proportion of markers were common to both maps and were used to construct the composite genetic linkage map. A novel approach to estimate recombination of two connected populations is described, where the meiosis information of all individuals is captured in a single estimator using a multipoint maximum likelihood estimation. The composite map consisted of 641 SNPs markers with a total map length of 1011 cM. This composite map represent the best consensus ordering of markers, a valuable reference framework for future studies in A. sellowiana. The large number of SNPs identified allowed us to construct high-density genetic maps, molecular tools which represent a relevant contribution for future genetic research and breeding efforts in A. sellowiana. / Acca sellowiana, conhecida como feijoa, pineapple guava ou goiabeira-serrana, é uma árvore frutífera nativa do Uruguai e do Brasil, pertencente a família Myrtaceae. A espécie destaca-se por suas frutas altamente aromáticas, com reconhecido valor nutracêutico e terapêutico. Apesar do promissor valor agronômico, os estudos genéticos nesta espécie são limitados. Os mapas genéticos são valiosas ferramentas em tais estudos, sendo empregados no desenvolvimento de estratégias de melhoramento molecular nos programas de melhoramento. No entanto, a falta de um elevado número de marcadores polimórficos nas populações de mapeamento é uma das principais limitações no desenvolvimento de mapas genéticos saturados. Recentemente, novos métodos de genotipagem baseados em tecnologia de sequenciamento de nova geração permitem identificar e genotipar milhares de marcadores em populações de mapeamento. Esta rápida e eficiente estratégia é muito útil para culturas pouco estudadas, com recursos genômicos limitados. Neste estudo, foram construídos mapas genéticos saturados em A. sellowiana. Foram usadas duas populações de mapeamento, H5 (\'TCO × BR\', n = 160) e H6 (\'TCO × DP\', n = 184), conectadas geneticamente pelo mesmo genitor feminino. A estratégia de genotipagem por sequenciamento (genotyping by sequencing; GBS) foi usada para simultaneamente identificar e genotipar marcadores de polimorfismos de nucleotídeo único (single nucleotide polymorphism; SNP) em ambas populações. No processo de detecção de SNPs, duas estratégias foram implementadas: na primeira foi empregado o genoma de referência de especie relacionada Eucalyptus grandis; na segunda, foi empregado uma abordagem de novo, que não requer genoma de referência. Após o processo de genotipagem quantitativo, 5350 e 4227 SNPs de alta qualidade foram selecionados para mapeamento em H5 e H6, respectivamente. Os mapas integrados H5 e H6 compreendem 1236 e 1302 marcadores distribuídos no 11 grupos de ligação esperados. Os mapas genético abrangeram um comprimento total de 1593 cM e 1572 cM, com uma distância média entre marcadores de 1:29 cM e 1:21 cM, nas populações H5 e H6, respectivamente. Um alto nível de colinearidade foi observado entre os dois mapas. Além disso, uma grande proporção de marcadores foram mapeados em ambos mapas e posteriormente usados na construção de um mapa genético integrado, considerando a informação de ambas populações simultaneamente. Foi apresentada uma nova abordagem para estimar as frações de recombinação em duas populações conectadas, onde a informação das meioses de todos os indivíduos é capturado num único estimador, usando uma estimativa de máxima verossimilhança multiponto. O mapa integrado composto contém 641 marcadores SNP com um comprimento total de 1011 cM. Este mapa representa o melhor ordenamento consenso de marcadores, sendo una valiosa referencia para futuros estudos nesta espécie. O grande número de SNPs identificados permitiu-nos a construção de mapas genéticos de alta densidade, ferramentas moleculares que representam uma contribuição relevante para futuras pesquisas genéticas e avanços no melhoramento genético em A. sellowiana.

Made available in DSpace on 2015-03-26T13:42:06Z (GMT). No. of bitstreams: 1 texto completo.pdf: 1107978 bytes, checksum: 5a845c5ae4a67c50d70037eeeeb7eed9 (MD5) Previous issue date: 2008-07-28 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / The genetic mapping facilitates the breeding work once one or more marks of the genotype can be associated to controlling genes of qualitative and quantitative characteristics (QTL). Genetic maps for several species have been built by different groups of researchers with different molecular markers and populations. A way to generate maps more saturated for those species would be the integration of the existent maps. The key to integrate different maps is the presence of common marks among them. Only when there are a minimum number of common marks among the different maps, these can be integrated. This way, the objective of this work was to develop a process of integration of genetic maps and to test the efficiency of this process. Data from the simulation of genome and populations were generated and analyzed. A important factor to obtain solid data in a mapping work is the sample or population size. Based in these simulated data it was evaluated the optimum population size to study the integration of genetic maps. To obtain and study the consensus maps, parental genomes and samples of co- dominant F2, dominant F2 and backcrosses populations were simulated. The generated samples had 100, 200, 400 individuals with 3 linkage groups each and 11 dominant and co-dominant molecular marks spaced by 5, 10 and 15 centiMorgans. 10 repetitions were accomplished by sample, five used to construct the consensus maps with analysis multilocus and other five without analysis multilocus. It was concluded that the obtention of the consensus maps becomes more efficient with the increase of the population size. A population size of 200 individuals would be enough to rescue the original information. To study the obtention of the effective integrated map samples of co-dominant F2 and backcrosses populations were simulated. The generated samples had 100, 200 and 400 individuals, 21 marks for linkage group and markers equidistant by 5 cM, in a total of four simulations for co-dominant F2 and four for backcrosses. Each simulated genome was fragmented in four new maps which three maps had eight markers and one had nine markers, each one of these maps containing four markers that are anchors among the four maps. These new maps were aligned, orderly and integrated and then compared with the original map. It was concluded that the process of ordainment and integration are efficient to obtain the integrated map. The population size exercises influence on the map and the distances measures among the marks. / O mapeamento genético facilita o trabalho de melhoramento uma vez que uma ou mais marcas do genótipo podem ser associadas a genes controladores de características qualitativas e quantitativas (QTL). Mapas genéticos para diversas espécies tem sido construídos por diferentes grupos de pesquisadores com diferentes tipos de marcadores e diferentes tipos de populações. Uma maneira de gerar mapas mais saturados para as diversas culturas seria a integração dos mapas genéticos já existentes. A chave para integração de mapas distintos é a presença de marcas comuns entre os mapas. Somente quando há um número mínimo de marcas comuns entre os diferentes mapas, estes podem ser integrados. Deste modo, o objetivo deste trabalho foi desenvolver um processo de integração de mapas genéticos e testar a eficiência do processo de integração. Para isso foram gerados e analisados dados a partir da simulação de genoma e de populações. Um dos fatores de fundamental importância para se obter dados consistentes em um trabalho de mapeamento é o tamanho da amostra ou população a ser trabalhada. Com base nestes dados simulados, avaliou-se o tamanho ótimo de populações para estudo de integração de mapas genéticos. Para estudo e obtenção dos mapas consenso foram simulados genomas parentais e amostras de populações F2 codominantes, F2 dominantes e de retrocruzamentos. As amostras geradas foram de tamanho 100, 200 e 400 indivíduos com três grupos de ligação cada e 11 marcas moleculares co-dominantes e dominantes espaçadas a 5, 10 e 15 centimorgans por grupo de ligação. Foram realizadas 10 repetições por amostra, sendo que destas, cinco foram utilizadas para construção de mapas consenso com análise multiloco e outras cinco sem análise multiloco. Concluiu-se que a obtenção dos mapas consenso se torna mais eficiente com aumento do tamanho da população. Um tamanho populacional de 200 indivíduos seria o suficiente para resgatar as informações originais. Para estudo da obtenção do mapa integrado efetivo foram simuladas F2 co-dominante e retrocruzamento com tamanhos de 100, 200 e 400 indivíduos, com 21 marcas por grupo de ligação e marcadores eqüidistantes a 5 cM, em um total de quatro simulações para F2 codominante e quatro para retrocruzamentos. Cada genoma simulado foi fragmentado em quatro novos mapas de modo que, foram obtidos três mapas com oito marcadores e um com nove marcadores, sendo que cada um deles constando quatro marcadores que são âncoras entre os quatro mapas. Estes novos mapas foram alinhados, ordenados, integrados e em seguida comparados com o mapa de origem. Concluiu-se que os processos de ordenamento e integração são eficientes para obtenção do mapa integrado e, também, que o tamanho da população exerce influência sobre o mapa e as medidas de distâncias entre as marcas.

Depuis plusieurs années, les programmes de recherche chez le cacaoyer ont mis l'accent sur l'étude des bases génétiques des caractères agronomiques d'intérêt, notamment concernant la résistance aux maladies et la qualité des fèves de cacao, qui représentent deux attributs importants pour la cacaoculture et la production de chocolat. Ce travail présente, par l'exploration du transcriptome et du génome du cacaoyer, la constitution de ressources moléculaires et l'analyse des voies de biosynthèse impliquées dans plusieurs de ces caractères agronomiques d'intérêt. L'étude du transcriptome a permis l'identification de plusieurs dizaines de milliers de gènes exprimés dans divers organes et pour différentes conditions environnementales et ont fourni de nombreux marqueurs moléculaires qui ont été utilisés pour réaliser des cartes génétiques haute densité. Les informations apportées par ce travail ont permis d’engager le séquençage du génome de la variété Criollo du cacaoyer. Son analyse et son annotation ont apporté un ensemble d'informations biologiques cruciales, depuis le catalogue des gènes et éléments mobiles jusqu'aux aspects évolutifs qui a révélé une structure du génome peu remanié par rapport à l'ancêtre commun aux dicotylédones. Par la suite, les travaux que nous avons menés pour améliorer la séquence complète ont conduit à une réduction considérable de la fragmentation chromosomique observée dans la première version. Par ailleurs 97% de la séquence assemblée et 99% des gènes sont désormais ancrés sur les chromosomes du cacaoyer. Pour commencer à exploiter cette nouvelle séquence du génome, nous avons conduit une étude QTL à partir d'une descendance entre Trinitario implantée en Guyane, permettant de localiser les régions génomiques impliquées dans la variation de la couleur des fèves de cacao. En s'appuyant sur la version améliorée du génome du Criollo, nous avons identifié deux gènes potentiellement impliqués dans la voie de biosynthèse des anthocyanines et flavonoïdes dans la principale région génomique concernée. Un des deux gènes, situé proche du marqueur situé au pic du QTL et codant pour une chalcone synthase, semble être un gène candidat prometteur. L'étude comparative de sa structure dans le génome du Criollo (à fève blanche) et du génome de l'Amelonado (à fève violette) a mis en évidence des différences structurales pouvant être à l'origine d'une modification fonctionnelle. L'ensemble des résultats présentés dans ce travail de thèse apporte une connaissance et des outils variés qui peuvent être exploités par de multiples approches intégrées pour étudier la génétique du cacaoyer. / For several years, cocoa research programs have focused on studying the genetic basis of agronomic traits of interest, especially for disease resistance and quality of cocoa beans, two important attributes for cocoa farmers and chocolate production. This work presents, by exploring the transcriptome and cocoa genome, the constitution of molecular resources and the analysis of biosynthesis pathways involved in several of these agronomical traits. The transcriptome study allowed the identification of several tens of thousands of genes expressed in various organs and for different environmental conditions and provided numerous molecular markers, used to produce high-density genetic maps. The information provided by this work led to the genome sequencing of a cocoa Criollo variety. Its analysis and annotation have provided a set of crucial biological information, from the catalog of genes and transposable elements to evolutionary aspects, which revealed a close evolutionary relationship to the eudicot putative ancestor, showing a limited number of recombination between ancestral chromosomes. Subsequently, the work we carried out to improve the complete sequence led to a considerable reduction of the chromosomal fragmentation observed in the first version. In addition, 97% of the assembled sequence and 99% of the genes are now anchored on the cocoa chromosomes. To exploit this new sequence of the genome, we conducted a QTL study from a progeny between Trinitario clones established in French Guiana, allowing to identify genomic regions involved in color trait variation observed in cocoa beans. Based on the improved version of the Criollo genome, we identified two genes potentially involved in the biosynthesis pathway of anthocyanins and flavonoids in the main genomic region concerned. One of the two genes is located nearby the QTL peak and encoding a chalcone synthase, appears to be a promising candidate gene. The comparative study of its structure in the Criollo genome (white bean) and in the Amelonado genome (purple bean) revealed several variations that could be responsible for functional modifications. The results presented in this thesis provide a variety of knowledge and tools useful to conduct multiple integrated approaches to studying cocoa tree genetics.

A narrow germplasm base and a complex allotetraploid genome have historically made the discovery of single nucleotide polymorphism (SNP) markers difficult in cotton (Gossypium hirsutum). We conducted a genome reduction experiment to identify SNPs from two accessions of G. hirsutum and two accessions of G. barbadense. Approximately 2 million sequence reads were assembled into contigs with an N50 of 508 bp and analyzed for SNPs. A total of 11,834 and 1,679 SNPs between the accessions G. hirsutum and G. barbadense, respectively, were identified with highly conservative parameters (a minimum read depth of 8x at each SNP and a 100% identity of all reads within an accession at the SNP). Additionally, 4,327 SNPs were identified between accessions of G. hirsutum in and assembly of Expressed Sequence Tags (ESTs). 320 and 252 KASPAR assays were designed for SNP mapping in non-genic and genic regions respectively. 187 markers in total (136 non-genic, 51 genic) were mapped using KBioscience KASPar genotyping assays in a segregating F2 population using the Fluidigm EP1 system. EST The target genome of EST markers was successfully predicted bioinformaticly diploid reference sequences. Examination of nucleotide substitutions and SNP frequencies further confirms validity of new markers. A genetic map was constructed using a large G. hirsutum segregating F2 population. Genetic maps generated by these newly identified markers will be used to locate quantitative, economically important regions within the cotton genome.

Powdery mildew, caused by the pathogen Blumeria graminis (DC) Speer (Syn. Erysiphe graminis DC) f. sp. tritici, is a major disease of wheat (Triticum aestivum L.). Race-specific resistance is easily identified in the field due to its qualitative phenotype and it is easy to incorporate because it is inherited as a single gene. Unfortunately, this type of resistance is easily overcome by the pathogen. Traits associated with quantitative trait loci (QTL) such as adult-plant resistance (APR), have become popular with plant breeders because of their durability over a wide geographic range and time. Due to the quantitative nature of these genes, they are difficult to study requiring multiple assessments of disease development under natural conditions in more than one location over a period of several weeks. Numerous QTL for APR to powdery mildew have been mapped in independent studies in different wheat backgrounds. The wheat cultivar Massey has been the subject of several studies due to its APR to powdery mildew that has remained effective for several decades. However, it has been difficult to identity simple sequence repeat (SSR) markers that are tightly linked to the QTL for APR in Massey. Such markers give breeders an advantage by allowing them to quickly identify and select for traits that would be difficult to distinguish in the field among breeding progeny from several backgrounds. Therefore, identification of tightly linked markers associated with APR to powdery mildew is necessary so that these traits can be selected for reliably in progeny. / Master of Science

O maracujá-doce (Passiflora alata Curtis) é uma espécie nativa no Brasil. Seu cultivo tem crescido nos últimos anos devido a sua valorização no mercado in natura e seus usos na fitoterapia. Entretanto, os cultivos comerciais enfrentam problemas devido a ocorrência da bacteriose (Xanthomonas axonopodis pv. passiflorae). O patógeno é endêmico no país, apresentando considerável variabilidade genética em suas populações naturais. Este trabalho teve como objetivo identificar QTL relacionados à resistência de P. alata à bacteriose em uma população F1 segregante oriunda do cruzamento entre acessos não endogâmicos. Foram avaliados os caracteres: área total da folha (TA), idade de queda da folha inoculada (IK), área total da lesão foliar (LA), área de clorose foliar (CA) e área da necrose foliar (NEA). Apenas um dos isolados apresentou diferenças de severidade em relação aos demais, sendo o menos agressivo (PA8-2). A inoculação do isolado M-129 mostrou que há variação significativa na resposta da população ao patógeno, sendo possível a identificação de genótipos transgressivos. A herdabilidade dos caracteres variou de 45% a 71%%. Foi construído um mapa de ligação integrado com 1.786 cM e uma densidade média de 4,5 cM. A análise de marcas individuais indicou a associação de 51 marcas aos fenótipos avaliados. O mapeamento de QTL, realizado por intervalo composto e utilizando uma estratégia diferenciada para populações F1 segregantes, identificou regiões associadas a 26 QTL para os cinco caracteres avaliados, sendo 16 deles referentes à LA, CA e NEA. A variação fenotípica explicada individualmente pelos marcadores variou de 0,8% a 16,7%. Sugere-se que a resistência à bacteriose é quantitativa, com predominância de efeitos genéticos aditivos. / The sweet passion fruit (Passiflora alata Curtis) is a specie native to Brazil. Its cultivation has increased in recent years due to its market valuation in natura and their uses in herbal medicine. However, crops are facing problems due to the occurrence of bacterial blight (Xanthomonas axonopodis pv. passiflorae). The pathogen is endemic in the country, with considerable genetic variability in their natural populations. This study aimed to identify QTL related to resistance of P. alata to bacterial blight in an F1 segregant population from the cross between outbred accessions. Five traits were evaluated: total area of the leaf (TA), age of inoculated leaf fall (IK), area of the leaf´s lesion (LA), area of the leaf´s chlorosis (CA) and area of the leaf´s necrosis (NEA). Only one of the isolates showed differences in severity in relation to others, being the least aggressive (PA8-2). The inoculation of the isolate M-129 showed significant variation in population response to the pathogen, making it possible to identify transgressive genotypes. The heritability of characters ranged from 45% to 71%. An integrated linkage map was constructed, with a length of 1,786 cM and an average density of 4.5 cM. The analysis of individual marks indicated the association of 51 markers to phenotypes. The QTL mapping was performed using composite interval and a special strategy for segregating F1 populations, identified 26 regions associated with QTL for the five traits, 16 of them related to LA, CA and NEA. The phenotypic variation explained by individual markers ranged from 0,8% to 16,7%. It is suggested that the resistance to bacterial blight is quantitative, with a predominance of additive genetic effects.