Submitted by Cláudia Bueno (claudiamoura18@gmail.com) on 2015-11-04T15:30:38Z No. of bitstreams: 2 Dissertação - Miriam Suzane Vidotti - 2015.pdf: 1671672 bytes, checksum: 08f31cfb7a6ed7c17a0a63b010b31e83 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2015-11-05T10:11:48Z (GMT) No. of bitstreams: 2 Dissertação - Miriam Suzane Vidotti - 2015.pdf: 1671672 bytes, checksum: 08f31cfb7a6ed7c17a0a63b010b31e83 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2015-11-05T10:11:48Z (GMT). No. of bitstreams: 2 Dissertação - Miriam Suzane Vidotti - 2015.pdf: 1671672 bytes, checksum: 08f31cfb7a6ed7c17a0a63b010b31e83 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2015-02-12 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / The upland rice has a high potential for expansion in the Brazilian savannahs region, where soils have limiting factors to yields, such as low phosphorus (P) availability. Considering the costs and the natural reserves scarcity, a suitable alternative is to explore the genetic resources to develop cultivars with P deficiency tolerance and P-use efficiency. Given the polygenic inheritance of these traits, a strategy is to concentrate the favorable alleles gradually through recurrent selection. The objectives of this study were to estimate genetic and phenotypic parameters, and evaluate progenies from CNA9/3/1 recurrent selection population of upland rice for the tolerance to P deficiency in the soil. In 2011/12 crop year, in Santo Antônio de Goiás-GO and Sinop-MT, two contrasting trials were conducted within each location: low and high P-level determined by the clay content of the soil, and without discrimination of other crop practices. The simple square lattice design 14 x 14 was adopted, with 189 S0:2 progenies and seven checks, with plots consisted of four 3-m-long rows. Grain yield (GY, in kg ha-1), plant height (PH, cm) and days-to-flowering (DF, in days) were evaluated. The individual variance analysis by location and by P-level were obtained, well as their phenotypic, genetic and environmental correlations. Only for PG, genetic and phenotypic parameters were estimated from joint analysis of variance within each P-level, and involving locations and P-levels. Estimates of the expected response to selection were obtained for the two P-levels, considering the direct, indirect and general selection criteria, and also the Harmonic Mean of Relative Performance (𝑀𝐻𝑃𝑅) and Tolerance Index (𝐼𝑇) of the genotypes identified as efficient and responsive from a graphic dispersion of adjusted means. The overall average of the GY was reduced significantly in 27% for P deficiency in the soil and the genetic variance among S0:2 progenies (𝜎̂𝑝 2) was narrowed in 63%. Highly significant diferences (p≤0.01) were observed among progenies, demonstrating the presence of sufficient genetic variability for GY, in this population, to obtain genetic progress in both P-levels. The progenies showed performance significantly (p≤0.01) below the of checks and also a significant interaction (p≤0.01) with locations, in both P-levels. The genotype x Plevel interaction was significant (p≤0.01) and the phenotypic correlation (√R2) between Plevels was low (0.44). In this sense, the expected response to selection demonstrated that progenies must be evaluated in P-deficient environments to obtain effective genetic gain in this environment. The genetic correlations between traits were significant (p≤0.01) and are favorable for the selection of genotypes with tolerance to P deficiency, short plant height and early cycle. However, there is a greater difficulty in increasing the yield potencial, since there the genetic correlation between this trait and PH was significant (p≤0.01) and positive (𝑟𝐺= 0.385). From the graphic dispersion of S0:2 progeny means have been identified 57 efficient and responsive to the presence of P in the soil, and for the selection within this group, the 𝑀𝐻𝑃𝑅 was more appropriate than 𝐼𝑇, because considering simultaneously these two attributes. / O arroz de terras altas possui um elevado potencial de expansão na região de cerrado brasileiro, onde os solos possuem fatores limitantes a produção, como a baixa disponibilidade de fósforo (P). Considerando os custos de produção e a escassez de recursos naturais, uma alternativa adequada é explorar os recursos genéticos afim de desenvolver cultivares tolerantes à deficiência e mais eficientes no uso do P. Tendo em vista a natureza poligênica desses caracteres, a seleção recorrente é uma alternativa que permite concentrar gradualmente os alelos favoráveis. O objetivo deste estudo foi estimar parâmetros genéticos e fenotípicos e avaliar progênies da população CNA9/3/1 de seleção recorrente de arroz de terras altas, quanto a tolerância à deficiência de P no solo. Em 2011/12, em Santo Antônio de Goiás-GO e em Sinop-MT, foram conduzidos dois experimentos contrastantes por local: baixo e alto nível de P, estabelecidos pelo teor de argila do solo, e sem a discriminação dos demais fatores. Foi adotado o delineamento látice quadrado simples 14 x 14, com 189 progênies S0:2 e sete testemunhas, com parcelas de 4 linhas de 3 m de comprimento. Foram avaliados os caracteres produção de grãos (PG, em kg ha-1), altura de planta (AP, em cm) e dias para o florescimento (DF, em dias). Foram obtidas as análises de variância individual por local e por nível de P para esses caracteres e as respectivas correlações fenotípicas, genéticas e ambientais. Somente para PG, foram realizadas as análises de variância conjunta dentro de cada nível de P e, posteriormente, envolvendo níveis e locais, foram estimados os parâmetros genéticos e fenotípicos. Foram obtidas as estimativas da resposta esperada com a seleção direta, indireta e geral considerando os dois níveis de P no solo, além da Média Harmônica da Performance Relativa (𝑀𝐻𝑃𝑅) e do Índice de Tolerância (IT) para os genótipos identificados como eficientes e responsivos a partir de uma distribuição gráfica de médias ajustadas. A deficiência de P no solo reduziu expressivamente a média geral em 27% para PG, além de estreitar a estimativa da variância genética entre progênies S0:2 (𝜎̂𝑝 2) em 63%. Foram observadas diferenças altamente significativas (p≤0,01) entre as progênies, evidenciando a presença de variabilidade genética suficiente na população para progresso genético em PG sob ambos os níveis de P. Essas apresentaram desempenho significativamente (p≤0,01) inferior ao das testemunhas e, também, interação significativa (p≤0,01) com locais, em ambos os níveis de P. Verificou-se que o efeito da interação de genótipo x nível de P foi significativo (p≤0,01) e a correlação fenotípica (√R2) entre níveis de P foi baixa (0,44). Nesse sentido, a resposta esperada com a seleção demonstrou que as progênies devem ser avaliadas em ambientes com deficiência de P para que o ganho genético nesse ambiente seja efetivo. As correlações genéticas entre os caracteres avaliados foram significativas (p≤0,01) e demonstraram ser favoráveis para a seleção de genótipos tolerantes à deficiência de P, porte baixo e ciclo precoce. Porém, existe maior dificuldade de se incrementar o potencial produtivo, visto que houve uma significativa (p≤0,05) correlação positiva entre esse caráter e AP ( 𝑟𝐺= 0,385). A partir da dispersão gráfica das progênies S0:2 foram identificadas 57 eficientes e responsivas à presença de P, sendo que para a seleção dentro desse grupo, 𝑀𝐻𝑃𝑅 demonstrou ser um critério mais adequado que 𝐼𝑇 ao se considerar simultaneamente esses dois atributos.

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2017-04-18T14:23:20Z No. of bitstreams: 2 Dissertação - Matheus Souza de Barros - 2015.pdf: 1949086 bytes, checksum: 889a6c45d4ce66700cc7ff366ef3caed (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-04-18T14:23:40Z (GMT) No. of bitstreams: 2 Dissertação - Matheus Souza de Barros - 2015.pdf: 1949086 bytes, checksum: 889a6c45d4ce66700cc7ff366ef3caed (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-04-18T14:23:40Z (GMT). No. of bitstreams: 2 Dissertação - Matheus Souza de Barros - 2015.pdf: 1949086 bytes, checksum: 889a6c45d4ce66700cc7ff366ef3caed (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2015-08-10 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The genetic improvement resulted from plant breeding acts decisively in maintaining the supply of agricultural foods like rice. The contribution of this genetic improvement to increase productivity is expressed in many crops by genetic progress which represents the genetic changes in the course of the selection cycles, and promotes the advance of genotypic average towards selection. Among the strategies adopted to increase the gain by selection, the early generation selection proves to be promising. This work has aimed to evaluate the genetic potential for selecting upland rice progenies and estimate the genetic progress for grain yield. The data used in this study were obtained from the progeny yield trials conducted in the period of eleven agricultural years 2002/03 to 2012/13 conducted by Embrapa Rice and Beans. The traits analyzed were grain yield (kg ha-1), plant height (cm) and days to flowering (day). In each year of the experiment, a group of progeny was tested in four to six sites. It was used Federer's augmented block design without replication per site in seven years and with at least three checks, in four years of the series were used two replications. The experimental data within each year were submitted to joint analysis. A mixedeffects linear model was applied for estimating the components of variance by the method of restricted maximum likelihood (REML). From this estimate of the components, it was calculated the genetic and phenotypic parameters, in addition the selective accuracy, the experimental precision coefficient, the experimental variation coefficient and relative variation coefficient. The genetic progress was estimated by the method of generalized linear regression of the adjusted means of progeny groups evaluated in each year by the mixed model approach. The estimates of the relative annual mean gain and the total relative gain for the three studied traits were also obtained. The estimates of genetic variance among progenies for grain yield were highly significant (p < 0.001), except for the progeny group evaluated in 2007. The heritability estimation ranging from 0.22 to 0.69 that, associated with the selective accuracy, indicates the expected level of efficiency with early generation selection in each group of progenies. For Plant height and daysto- flowering, the genetic variance estimates were significant (p ≤ 0.01) in all groups. The variance components for these traits led to a rather high heritability estimation that suggests favorable conditions for selection in early generations. The genetic progress for grain yield (80.5 kg ha-1 yr-1) was highly significant. This value represents a relative annual mean gain of 2.88%. Throughout the period, the cumulative gain was estimated in 32.86% which indicates an increase of 918 kg ha-1 for grain yield. The response for plant height was not significant, suggesting that the height of progenies remained stable over the period. For days-to-flowering, it was detected significant increase in cycle length, indicated by the cumulative increase of the vegetative period in about five days (6.73%). From these results we conclude that early generation selection, adopted by the breeding program, were effective in promoting the genetic gain for grain yield in the elite populations. / O melhoramento genético de plantas atua de modo decisivo na manutenção da oferta de alimentos de origem agrícola como o arroz. A contribuição do melhoramento para o aumento da produtividade em várias culturas é expressa pelo progresso genético, que representa as alterações genéticas, no decorrer dos ciclos seletivos, promovendo o deslocamento da média genotípica do caráter no sentido da seleção. Entre várias estratégias adotadas para aumentar os ganhos com a seleção merece destaque a seleção precoce, que envolve a avaliação de progênies endogâmicas nas gerações F3 ou F4. Este trabalho teve como objetivo avaliar o potencial genético para a seleção em progênies de arroz de terras altas e estimar o progresso genético para produção de grãos. Foram utilizados dados dos ensaios de rendimento de progênies conduzidos no período de onze anos agrícolas, 2002/03 a 2012/13 pela Embrapa Arroz e Feijão. Foram analisados os caracteres produção de grãos (kg.ha-1), altura de plantas (cm) e dias para a floração (dia). Em cada ano da série um grupo de progênies foi testado em experimentos instalados entre quatro a seis locais. Foi empregado o delineamento blocos aumentados de Federer em sete anos da série sem repetição por local e com no mínimo três testemunhas, e em quatro anos os ensaios foram duplicados. Os dados foram submetidos à análise conjunta dos experimentos dentro de cada ano. Por meio de um modelo linear de efeitos mistos foram estimados os componentes de variância pelo método da máxima verossimilhança restrita ou residual (REML). A partir da estimativa dos componentes foram calculados os parâmetros genéticos e fenotípicos. O progresso genético foi estimado pelo método da regressão linear generalizada das médias ajustadas dos grupos de progênies avaliadas em cada ano, pela abordagem de modelos mistos. Foram obtidas as estimativas do ganho médio relativo anual e do ganho relativo total. As estimativas de variância genética, para produção de grãos, entre progênies foram altamente significativas (p<0,001), exceto pelo grupo de progênie avaliado em 2007. As estimativas de herdabilidade variaram de 0,22 a 0,69, que associadas às estimativas de acurácia seletiva indicam o nível de eficiência esperado com a seleção precoce em cada grupo de progênies. Para os caracteres altura de plantas e dias para floração as estimativas de variância genética foram significativas (p≤0,01) em todos os grupos. Os componentes de variância para esses caracteres conduziram a estimativas de herdabilidade bastante elevadas, sugerindo, por tanto, condições favoráveis para seleção em gerações iniciais. O progresso genético estimado para produção de grãos (80,5 kg.ha-1ano-1) foi altamente significativo, representando um ganho relativo médio anual de 2,88%. Em todo o período o ganho acumulado estimado foi de 32,86%, que equivale ao incremento em produtividade de 918 kg ha-1. A resposta para altura de plantas não foi significativa, sugerindo que a estatura das progênies avaliadas permaneceu estável ao longo do período. Para dias para floração houve aumento significativo na duração do ciclo, expresso pelo acréscimo acumulado de aproximadamente cinco dias (6,73%) na duração do ciclo vegetativo. Conclui-se que a estratégia de seleção precoce, adotada pelo programa de melhoramento, foi eficiente em promover o progresso genético para o caráter produção de grãos nas populações elites.

No Brasil, a ferrugem, causada pelo fungo Puccinia psidii Winter, se destaca como a mais importante doença da cultura do eucalipto, principalmente na região Sudeste do país. O uso de genótipos resistentes é o principal método de controle da doença, o que torna essencial o conhecimento das bases genéticas da resistência para a seleção de genótipos resistentes. Neste estudo, uma progênie F1 de 90 irmãos completos foi utilizada para identificar marcadores ligados a genes de resistência a este patógeno. A severidade da doença foi avaliada em quatro épocas ao longo de uma epidemia por meio de escala diagramática e os dados foram utilizados para o cálculo da área sob a curva de progresso da doença (AUDPC) para cada indivíduo. O delineamento experimental foi de blocos incompletos em dois ambientes e os valores de AUDPC foram analisados por meio de modelos mistos, excluindo-se a variação ambiental. Um mapa genético foi construído com base em marcadores microssatélites, AFLP e TRAP através de uma análise multiponto. O mapeamento de QTL seguiu duas abordagens: mapeamento por intervalos (IM) e mapeamento por intervalos compostos (CIM). No mapa de ligação foram posicionados 117 marcadores microssatélites, 10 TRAP e 33 AFLP, distribuídos por 11 grupos de ligação, totalizando 1075 cM com distância média de 6,7 cM entre marcadores. A análise de IM identificou um QTL (LOD=7,7) no grupo de ligação 3, representando 28,5% da variação em AUDPC na progênie observada. Já a análise CIM detectou dois QTL, ambos no grupo de ligação 3. A posição do primeiro QTL coincidiu com aquela do QTL identificado na análise por IM, porém com maior valor de LOD (10,3), explicando 39,5% da variação em AUDPC. O segundo QTL (LOD=3,4), cujo valor probabilístico máximo distou 39,0 cM do primeiro QTL, explicou 6,9% da variação fenotípica. Os alelos de resistência nos dois QTL se encontram em repulsão, com efeito aditivo negativo (=-0,60) e ausência de dominância no primeiro QTL e efeito aditivo positivo (=0,29) e dominância completa (=-0,23) no segundo QTL. Os marcadores ligados aos dois QTLs têm potencial utilização na seleção assistida, com conseqüente aumento na eficiência durante a seleção de genótipos resistentes. / In Brazil, the eucalyptus rust, caused by Puccinia psidii Winter, stands out as the most important disease of eucalypts, mainly in southeastern Brazil. The use of resistant genotypes is the main control method, which makes the understanding of the genetic basis of resistance essential to the selection of resistant genotypes. In this study, an F1 progeny of 90 plants was used to identify markers linked to resistance genes to this pathogen. Four disease severity assessments were conducted during an epidemic with a diagrammatic scale and the data were used to calculate the area under the disease progress curve (AUDPC) for each plant. The experimental design consisted of incomplete blocks in two environments. AUDPC values were analyzed using mixed models, excluding the environmental variation between locations. A genetic map was constructed based on microsatellite, AFLP and TRAP markers based on a multipoint analysis. QTL mapping was done based on two approaches: interval mapping (IM) and composite interval mapping (CIM). The linkage map consisted of 117 microsatellite, 10 AFLP and 33 TRAP markers, placed over 11 linkage groups, totalizing 1075 cM with an average distance of 6.7 cM between markers. IM analysis identified a QTL (LOD = 7.7) on linkage group 3, accounting for 28.5% of the phenotypic variation in AUDPC. The CIM analysis detected two QTL, both on linkage group 3. The position of the first QTL coincided with that of the QTL identified in the IM analysis, but with a higher LOD (10.3), accounting for 39.5% of the variation. The second QTL (LOD = 3.4), whose maximum probability value was placed 39.0 cM away from the first QTL, accounted for 6.9% of the phenotypic variation. The resistance alleles in the two QTL are linked in repulsion with negative additive effect ( =- 0.60) and lack of dominance in the first QTL and positive additive effects (=0.29) and complete dominance (=-0.23) in the second. The markers linked to the two QTL have potential use in assisted selection, thus increasing the efficiency of selection of resistant genotypes.

Abstract Animals live in a heterogeneous world where threats and abundance and quality of resources vary across space and time. Heterogeneity induces uncertainty in decisions that animals must make, e.g., where to breed. Adaptive decisions may be facilitated by personally collecting information on the quality of the environment and by observing the behaviour and success of other individuals. Such social information use is common in nature. I investigate information use in relation to ecological threats (brood parasites, nest predators) and long-term information use in breeding site choice in the wild. Moreover, I examine the genetic basis of social cue use in breeding site choice. I demonstrated experimentally that open-nesting hosts of a brood parasite, the common cuckoo (Cuculus canorus), can cue on cuckoo vocalizations to estimate cuckoo abundance and avoid breeding sites with high perceived parasitism risk. Another experiment showed that pied flycatchers (Ficedula hypoleuca) derive predation risk information from the fates of heterospecific nests, can associate the information with a nest site characteristic and generalize the association to own nest site choice. However, apparently only young females that made their choice quickly used the information in nest site choice. Pied flycatchers were further observed to collect habitat quality information based on the old nest contents during the post-breeding period. Use of the information in breeding site choice in the following spring varied between sex and age groups as well as geographically. Some birds integrated the post-breeding period information with the information available during settlement suggesting sequential social information use. Finally, quantitative genetic analyses revealed low additive genetic variances and genetic heritabilities of social cue use in breeding site choice in a collared flycatcher (Ficedula albicollis) population. These results demonstrate new aspects of informed habitat choice in wild animals which have important implications for species coexistence and community ecology, parasite-host coevolution, between-species niche dynamics and evolution. Between-individual variation in information use is highlighted throughout the thesis and warrants further research. The evolutionary potential of information use appears low, but more studies in other populations and species are needed. / Tiivistelmä Eläimet elävät ympäristössä, jossa resurssit ja uhat vaihtelevat ajallisesti ja alueellisesti. Tämä vaihtelu aiheuttaa epävarmuutta eläinten päätöksentekoon, kuten pesimäpaikan valintaan. Hyödyllisten päätösten tekoa voi edesauttaa keräämällä tietoa ympäristön laadusta itsenäisesti tai seuraamalla muiden yksilöiden käytöstä ja menestystä. Tällainen sosiaalisen informaation käyttö on yleistä eläinkunnassa. Tutkin informaation käyttöä ekologisten uhkien (pesäloiset, -pedot) suhteen ja pitkäaikaista informaation käyttöä pesimäpaikan valinnassa luonnonpopulaatioissa. Lisäksi selvitän pesimäpaikan valintaan liittyvän informaation käyton geneettistä periytyvyyttä. Selvitin kokeellisesti, että pesäloisen, käen (Cuculus canorus), isäntälajit voivat käyttää käkien ääntelyä vihjeenä alueellisesta loisintauhasta ja siten välttää korkean uhan alueita pesimäpaikan valinnassa. Toisessa kokeessa havaittiin kirjosieppojen (Ficedula hypoleuca) keräävän tietoa pesäpetouhasta toisen lajin pesätuhojen kautta, kykenevän yhdistämään tiedon erilliseen pesäpaikan ominaisuuteen ja käyttämään tätä assosiaatiota omassa pesäpaikan valinnassa. Kuitenkin vain nuoret naaraat, jotka tekivät valintansa nopeasti, käyttivät kyseistä informaatiota valinnassaan. Lisäksi havaitsin kirjosieppojen keräävän tietoa ympäristön laadusta pesinnän jälkeen vanhojen pesäsisältöjen avulla. Kyseisen tiedon käyttö pesimäpaikan valinnassa seuraavana keväänä vaihteli lintujen sukupuolen ja iän suhteen, kuin myös alueellisesti. Osa linnuista yhdisti pesimäpaikan valinnassaan aikaisempaa, pesinnän jälkeen kerättyä tietoa ja keväällä saatavilla olevaa sosiaalista informaatiota. Geneettinen analyysi viittasi pesimäpaikan valintaan liittyvän informaation käytön alhaiseen additiivisen geneettisen varianssin määrään ja siten alhaiseen geneettiseen periytyvyyteen sepelsiepolla (Ficedula albicollis). Väitöskirjani tulokset kuvaavat uudenlaisia informaation käytön muotoja eläinten pesimäpaikan valinnassa. Havainnot auttavat ymmärtämään pesälois-isäntä rinnakkaisevoluutiota, lajien välisiä vuorovaikutuksia, lajiyhteisöjen toimintaa ja evoluutiota. Yksilöiden välinen vaihtelu informaation käytössä näyttää olevan yleistä, ja lisätutkimuksen tarpeessa. Informaation käytön evolutiivinen potentiaali näyttää rajalliselta, mutta lisätutkimukset eri populaatioilla ja lajeilla ovat tarpeen.

Comme ils conduisent à des modifications de sensibilité toxicologique et de traits d'histoire de vie au sein des populations naturelles exposées, les phénomènes d'adaptation à la contamination questionnent la pertinence de l'évaluation du risque environnemental lié au rejet de contaminants chimiques et sont devenus un champ de recherche important à développer en écotoxicologie. Centrés sur l'étude des capacités d'adaptation de l'amphipode d'eau douce Gammarus fossarum, les travaux présentés ici ont eu pour objectif d'avancer dans la compréhension des effets d'une exposition à long terme au cadmium à l'échelle populationnelle et en milieu naturel. En recourant à des méthodologies de biomonitoring par encagement, de culture et d'exposition au laboratoire et de suivi démographique in situ, une démarche couplant études a priori et rétrospectives sur populations naturelles, a permis d'identifier un phénomène d'augmentation de la tolérance et de modification de traits d'histoire de vie chez une population exposée historiquement au cadmium. Des approches de génétique quantitative conduites sur trois populations ont dans un deuxième temps mis en évidence (1) une faible héritabilité de la sensibilité au cadmium au sein de populations naïves ; et (2) un rôle majeur des effets parentaux induits par l'exposition comme mécanisme populationnel soutenant l'évolution de la tolérance chez cette espèce. Enfin, la caractérisation de la variabilité de la sensibilité au cadmium au sein du genre Gammarus (dix-sept populations) et la mise en regard de la divergence de la population tolérante vis-à-vis de cette variabilité, ont permis de discuter des implications de ces processus évolutifs induits par des expositions environnementales pour l'évaluation du risque lié aux substances chimiques. Ainsi, il apparaît nécessaire de prendre en compte les réponses adaptatives induites par la contamination des milieux comme source de variabilité et d'incertitude afin de proposer une évaluation du risque pertinente intégrant pleinement l'ensemble des impacts des contaminations environnementales sur les populations naturelles / Because they lead to changes in toxicological sensitivities and life history traits within field populations, evolutionary processes supporting adaptation to contamination challenge the relevance of environmental risk assessment of chemical contaminants. Hence their study becomes an important developing field of research in ecotoxicology. Focusing on the study of adaptive capacity of the freshwater amphipod Gammarus fossarum, this work aimed to gain insight into the effects of long term exposure to cadmium at the population scale in the field. By means of biomonitoring methodologies (caging), population demographic sampling, culture and exposure in the laboratory we identified a phenomenon of increased tolerance and modification of life history traits in a natural population historically exposed to cadmium. Quantitative genetics experiments conducted on three populations secondly demonstrated (1) a low heritability of sensitivity to cadmium in naïve populations ; and (2) a major role of parental effects induced by exposure as populational mechanism supporting the development of tolerance in this species. Finally, the characterization of the variability of cadmium sensitivity in the genus Gammarus (seventeen populations), and the analysis of the divergence of the tolerant population in comparison to this variability led us to discuss about the implications of these evolutionary processes induced by environmental exposure for risk assessment of chemicals. Thus, from these results it appears necessary to take into account adaptive responses induced by environmental contamination as a source of variability and uncertainty in order to provide a relevant risk assessment fully integrating all the impacts of environmental contamination on natural populations

L’adaptation d’un champignon pathogène à son milieu, ainsi que l’évolution et la structuration des populations qui en découlent, sont fortement influencés par ses traits d’histoire de vie qui conditionnent sa fitness. C’est ce que nous avons illustré chez Melampsora larici-populina, l’agent de la rouille du peuplier. Ainsi, nous avons mis en évidence que le volume des spores du champignon évolue de manière répétable au cours des épidémies annuelles dans la vallée de la Durance, et ce sous l’effet de la sélection naturelle, signe que ce trait intervient directement dans le processus adaptatif du champignon. Par conséquent, les contraintes génétiques conditionnant le potentiel adaptatif du champignon en lien avec les traits d’histoire de vie ont été étudiées en laboratoire, au sein d’une descendance S1 issue de l’autofécondation d’une souche de référence. Les résultats obtenus suggèrent que M. larici-populina présente un potentiel adaptatif élevé. Enfin, une carte génétique à haute résolution du champignon, comprenant 18 chromosomes, a été construite afin d’étudier le déterminisme génétique de ces traits d'histoire de vie. Un locus de virulence ainsi que des QTL intervenant dans l’expression de la taille des lésions ont pu être détectés et positionnés avec précision sur cette carte. Ce travail a mis en évidence le rôle des traits quantitatifs dans l’adaptation et la structuration des populations de M. larici-populina en réponse aux pressions de sélection du milieu, en lui conférant un potentiel adaptatif élevée, essence de l’adaptation des organismes. Il ouvre également de nombreuses perspectives de recherche visant à identifier les bases génétiques de l’adaptation de ce champignon pathogène à son hôte, éléments indispensables à l’élaboration de stratégies de lutte durables / Adaptation of a phytopathogenic fungus to its environment, as well as the resulting evolution and structuration of its populations, are strongly influenced by its life history traits which condition its fitness. This is illustrated here with the poplar rust fungus Melampsora larici-populina. Hence, we showed that spore volume repeatedly evolved through natural selection, during annual epidemics in the Durance River valley, showing the implication of this trait in the fungus adaptive processes. Consequently, genetic constraints conditioning the adaptive potential of the fungus, in connection with life history traits, were studied in laboratory, over a progeny resulting from a selfing of a reference strain. Results suggest that M. larici-populina has a high adaptive potential. Finally, a high resolution genetic map of the fungus, comprising 18 chromosomes, has been built in order to study genetic determinism of these traits. One locus of virulence and three QTL involved in the expression of the lesion size were detected and accurately mapped on this map. This work emphasizes the role of quantitative traits in adaptation and structuration of M. larici-populina populations in response to the environmental selective pressures, by conferring an adaptive potential, the basis of organisms’ adaptation. It also opens many opportunities to identify the genetic bases of adaptation of this fungus, these elements being essential for the development of sustainable strategies of disease control

Les invasions biologiques sont un composant du changement global et ont des impacts dramatiques sur les écosystèmes, les agrosystèmes et la santé humaine. Néanmoins, ces processus biologiques particuliers offrent la possibilité d'étudier l'évolution phénotypique et génétique en un sur des temps écologiques. En effet, les invasions biologiques impliquent de fortes contraintes environnementales et démographiques sur les populations, et de forts effets sous-jacents de la sélection et de la dérive. Pourtant, les espèces envahissantes sont parmi les colonisateurs les plus prolifiques de la nature, et surprennent par leur capacité à répondre à ces contraintes. Le potentiel évolutif et adaptatif des populations envahissantes a été à de nombreuses reprises proposé comme facteur facilitant le succès de ces invasions. Qu'il s'agisse de processus génétiques d'adaptations (i.e. des changements de fréquences d'allèles) ou plastiques (i.e. un ajustement par plasticité phénotypique en réponse à un stimulus environnemental), la capacité de réponse à la sélection des espèces envahissantes les placent au centre des études de la biologie évolutive moderne. Ici, nous utilisons la récente invasion mondiale de la drosophile à ailes tachetées Drosophila suzukii pour étudier en détail les mécanismes de la réponse à la sélection potentiellement impliqués dans le succès de cette invasion. L'analyse des patrons de variations moléculaires neutre, nous ont permit de retracer l'histoire complexe de cette invasion mondiale, et d'évaluer par la suite la divergence phénotypique et l'évolution de la variation génétique quantitative en comparant les populations ancestrales de D. suzukii à leurs populations dérivées. Nous avons pu ainsi estimer les effets de la sélection et de la dérive génétique au cours de cette invasion, et discuter leur importance au regard de l'évolution de la forme de l'aile dans cette espèce. Enfin des protocoles expérimentaux d'analyse de la plasticité phénotypique ainsi que des méthodes de modélisation de niche climatique nous permettent de discuter l'influence de la fluctuation des conditions environnementales sur le succès de cette invasion. / Biological invasions are a component of global change and have dramatic effects on ecosystems, agrosystems and human health. Nonetheless, these peculiar biological processes offer a great opportunity for the study of rapid phenotypic and genetic evolution, at an ecological timescale. Biological invasions often involve environmental and demographic constraints on populations, as well as strong effects of selection and drift. However, these species are among the most successful colonialists in nature, and their ability to respond to these constraints is remarkable. The evolutionary and adaptive potential of invasive populations have been proposed as facilitating factors of the success of invasions. Processes of genetic (i.e. changes in allele frequencies) and plastic (i.e. adjustment to environmental fluctuation through phenotypic plasticity) involved in the success of biological invasions are at the center of modern evolutionary biology. Here, we use the recent spread of the spotted-wing Drosophila suzukii to study the underlying mechanisms of response to selection potentially involved in the success of this global invasion. Analyzing patterns of neutral genetic variation allowed us to decipher the complex history of this worldwide invasion, and subsequently evaluate phenotypic divergence and evolution of quantitative genetic variation among ancestral and derived populations. We thus estimated the effects of selection and drift throughout this invasion and discuss their importance regarding the evolution of wing shape in this species. Finally, experimental protocols on the analysis of phenotypic plasticity as well as Species Distribution Modeling methods allowed us to discuss the influence of environmental fluctuations on the success of this invasion.

Les interactions entre les contraintes environnementales sur le phénotype rendent complexe l'étude des mécanismes d'adaptation et d'évolution des plantes. Aborder une telle question nécessite une approche intégrée de l'orchestration, à différentes échelles, des réponses des plantes à des stimuli environnementaux isolés et combinés. Nous avons utilisé Arabidopsis thaliana pour évaluer les bases génétiques des réponses à deux contraintes abiotiques majeures en forte interaction au champ : la sécheresse et les hautes températures. Grâce à des outils performants pour l'analyse du phénotype, une large gamme de traits a été mesurée sur de nombreux génotypes différant dans leur plasticité. Nous avons caractérisé la croissance et des traits d'histoire de vie, la morphologie et la structure foliaire, ainsi que les capacités d'acquisition et de conservation des ressources, en particulier l'eau et le carbone. Après une description des réponses phénotypiques et de leur architecture génétique, les conséquences de ces réponses dans une perspective écologique et évolutive ont été évaluées. En particulier, nous avons analysé les variations des stratégies fonctionnelles mises en relief par les covariations phénotypiques et génétiques. De plus, les processus responsables des réponses observées ont été examinés. Les résultats ont permis de mettre en évidence des variations génétiques associées à plusieurs régions génomiques dont l'influence est probablement majeure pour les mécanismes d'adaptation des plantes. Certaines de ces régions génomiques, de par leur effet sur la performance, sont porteuses d'intérêt dans une perspective d'amélioration des espèces cultivées face aux changements climatiques actuels. / The mechanisms of plant adaptation and evolution are difficult to investigate since environmental constraints have interactive effects on plant phenotypes. Such study requires an integrated approach about the coordination, at different organizational levels, of the plant phenotypic responses to multiple environmental cues. Using the model plant Arabidopsis thaliana, we assessed the genetic bases of the integrated responses to two major abiotic constraints that strongly interact in the field: water availability and high temperature. Using powerful tools for the analysis of the phenotype, a large range of traits was measured in many genotypes that differ in their plasticity. We focused on the traits related to plant growth and life history, leaf structure and morphology, and to the acquisition and conservation of resources, specifically water and carbon. After a description of the phenotypic responses and their genetic architecture, the ecological and evolutionary consequences of these responses were evaluated. Specifically, we examined the variations in the functional strategies that are highlighted by phenotypic and genetic covariations. Moreover, the processes responsible of the observed phenotypic responses to environmental constraints were investigated. Strong genetic variability associated to particular genomic regions was identified. Such loci have presumably important influence on the mechanisms of plant adaption to fluctuating environments. Some of these genomic regions have a strong effect on plant performance in stressing conditions, and therefore offer promising avenues for crop improvement facing current global climate change.

Dans un contexte de réduction des intrants agricoles, la création de variétés de blé qui utilisent l’azote de manière plus efficiente est aujourd’hui nécessaire. Cette thèse, issue d'un partenariat public-privé entre l'Institut National de la Recherche Agronomique et Biogemma, avait pour but d'apporter des outils nécessaires à la création de variétés répondant à cette exigence. Pour ce faire, nous avons analysé 225 variétés commerciales génotypées avec 24K SNP et testées dans huit combinaisons d’année, lieu et régime azoté. Nous avons montré que même si la sélection a amélioré l’efficience d’utilisation de l’azote en condition optimale et sub-optimale, ce progrès génétique doit être accéléré et mieux réparti entre les différents traits. Nous proposons pour cela de mixer sélection phénotypique et sélection assistée par marqueurs. Dans ce sens, nous avons développé une méthode pour définir les régions chromosomiques associées à nos 28 traits. Parmi les 333 régions identifiées, nous avons notamment localisé le gène NAM-A1 et avons pu caractériser ses variants naturels. Nous avons aussi montré que la sélection génomique pourrait être plus efficace si les SNP étaient présélectionnés en fonction de leurs significativités en génétique d’association multi-environnementale. Les réseaux d’interactions épistatiques furent aussi étudiés, mettant en évidence un sous-réseau particulièrement intéressant. Nos résultats et méthodes sont discutés au regard des stratégies d’amélioration variétale et de découverte de gènes. Des pistes de recherche complémentaires et des améliorations ont aussi été suggérées. / In a context of fertiliser reduction, breeding for enhanced nitrogen use efficiency in bread wheat is necessary. This PhD thesis resulting from private-public collaboration between the French National Institute for Agricultural Research and Biogemma aimed providing necessary tools. Analyses were conducted using a dataset of 225 commercial varieties genotyped with 24K SNP and tested in eight combinations of year, location, and nitrogen regimes. We showed that even if past selection increased nitrogen use efficiency at high and moderate nitrogen regimes, genetic progresses need to be accelerated and better balanced between traits. This could be achieved by mixing phenotypic and marker assisted selections. In this sense, we developed a method to define quantitative trait locus from genome-wide association study: 333 chromosomal regions involved in 28 NUE-related traits have been identified. The NAM-A1 gene was located in one of these regions and its natural variants were characterized. We also showed that genomic selection could be improved by pre-selecting SNP based on their significance in a multi-environmental genome-wide association study. Networks of epistasis interactions were also studied and an interesting sub-network was identified. Results and methods are discussed regarding breeding and gene discovery strategy. Further investigations and improvements are suggested.

La production de perle de culture par l’huître perlière Pinctada margaritifera représente la seconde ressource économique après le tourisme en Polynésie Française. L’une des voies d’amélioration privilégiée de la qualité de production passe par la voie de la sélection génétique. Dans ce contexte, le déterminisme génétique des caractères d’intérêt perlicole et leurs variations à différentes échelles phénotypiques a été étudié. Les rôles respectifs de l’huître donneuse et de la receveuse, au travers de greffes expérimentales, ont révélées une corrélation positive entre les paramètres de croissance des coquilles de receveuses et la taille des perles, ainsi qu’un effet donneuse sur la qualité de la perle. Les analyses d'expressions de huit biomarqueurs de biominéralisation, codant des protéines des couches aragonitiques ou prismatiques, ont révélé une corrélation entre l’expression de ces gènes au niveau du sac perlier avec à la fois les paramètres de qualité des perles et de croissance des receveuses. L’âge de l’huître donneuse de greffon semble jouer un rôle déterminant aussi bien au niveau des phénotypes de la taille que pour le grade et les défauts de surface de la perle. Enfin, les valeurs d'héritabilité des phénotypes ont été estimées pour la première fois chez l'espèce, au travers de modèle animal utilisant des familles produites en écloserie. Globalement, les résultats montrent une transmission génétique linéaire et schématiquement on peut dire que les receveuses contrôlent principalement la croissance et la taille des perles, alors que les donneuses influencent leur qualité. / Cultured pearl production in the pearl oyster Pinctada margaritifera represents the second largest source of revenue after tourism, and it is the top export industry in French Polynesia. One of pearl farming industry’s greatest challenges is to “produce fewer but better pearls” through genetic improvement. To address this challenge, the genetic determinism of pearl quality traits and how they vary at different phenotypic scales was studied. The respective roles of donors and recipients was explored through uniform experimental grafts and revealed a positive correlation between recipient shell biometric parameters and pearl size, and a donor effect on pearl quality traits. Gene expression analyses of 8 biomineralisation biomarkers, encoding aragonitic and prismatic proteins, highlights a correlation between pearl sac gene expression with pearl quality traits and recipient shell biometric parameters. The age of the donor oyster also played a determining role with respect to the size phenotype and for grade and surface defects of the pearl. Finally, the heritability values of the phenotype were estimated for the first time for this species using an animal model on a family produced in a hatchery setting. Results show a linear genetic transmission and overall, suggest that the recipient oyster primarily controls the size of the pearls while the donor oyster influences the quality.