Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2015-05-26T13:09:56Z No. of bitstreams: 2 Dissertação - Patrícia Aparecida de Carvalho Felisberto - 2015.pdf: 5137723 bytes, checksum: 1a2d107b747e1e556b5cc29325451197 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2015-05-26T13:14:03Z (GMT) No. of bitstreams: 2 Dissertação - Patrícia Aparecida de Carvalho Felisberto - 2015.pdf: 5137723 bytes, checksum: 1a2d107b747e1e556b5cc29325451197 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2015-05-26T13:14:03Z (GMT). No. of bitstreams: 2 Dissertação - Patrícia Aparecida de Carvalho Felisberto - 2015.pdf: 5137723 bytes, checksum: 1a2d107b747e1e556b5cc29325451197 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2015-02-23 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / In the field aerial application of sugarcane ripeners, spraying should reach its target without drifting occurs, avoiding contamination of surrounding crops. The aim of this research was to evaluate the effects of sugarcane ripeners simulated drift using sublethal rates in maize when applied to V4 and V8 stages of development. For this purpose were installed two trials on the Federal University of Goiás, Regional Jataí. Each plot contained 10 rows of SYN 7G17 maize hybrid. Each trial consisted of eight treatments established in design with randomized blocks where the treatments were arranged at factorial design of 2 x 4, plus one control treatment, with four replications. First factor corresponds to spraying at the V4 and V8 stages of maize and the second factor to sublethal rates of 3; 6; 9 to 12% of 475.2 g a.i. ha-1 ethephon (Trial I), 216.0 g a.e. ha-1 glyphosate (Trial II) and 15.0 g a.i. ha-1 sulfometuron methyl (Trial III). At 20 and 36 days after sowing, when the maize were in V4 and V8, respectively, applications of ripeners were performed.The phytotoxicity effects were noticeable in both stages at 7, 14, 21, 28 and 42 days after spraying only for Trial III. For glyphosate, the length and diameter of ear were reduced in size to both stages. Maize is sensitive to sublethal rates starting from 0.45 g a.i. ha-1 sulfometuron methyl in V4 and V8 stages of development. The lower grain yield was only to plants subjected to sublethal doses of sulfometuron methyl, being lower yield when drift occurs in the V8 stage of development compared to the V4 stage. / Na aplicação aérea de maturadores de cana-de-açúcar, a pulverização deve atingir o seu alvo sem que ocorra deriva, evitando a contaminação de culturas circunvizinhas. Objetivou-se com a pesquisa avaliar o efeito de subdoses de maturadores de cana-de-açúcar na cultura do milho quando aplicadas em estádios de desenvolvimento V4 e V8. Foram instalados três ensaios na Fazenda da Universidade Federal de Goiás, UFG, Regional Jataí. Cada parcela continha 10 linhas do híbrido de milho SYN 7G17. Cada ensaio foi constituído de oito tratamentos estabelecidos em delineamento com blocos casualizados em esquema fatorial 2 x 4 mais um tratamento controle (sem aplicação de maturador), com quatro repetições. O primeiro fator corresponde a aplicação nos estádios V4 e V8 da cultura do milho e o segundo fator às subdoses de 3; 6; 9 e 12% de 475,2 g i.a. ha-1 de ethephon (Ensaio I), 216,0 g e.a. ha-1 de glyphosate (Ensaio II) e 15,0 g i.a. ha-1 de sulfometuron methyl (Ensaio III). Aos 20 e aos 36 dias após a semeadura, quando as plantas de milho se encontravam em V4 e V8, respectivamente, foram realizadas as aplicações dos maturadores. Os efeitos de fitotoxicidade foram perceptíveis em ambos os estádios aos 7, 14, 21, 28 e 42 dias após a aplicação apenas para o Ensaio III. Para glyphosate, o comprimento e diâmetro de espigas foram afetados para ambos os estádios. A cultura do milho é sensível a subdoses a partir de 0,45 g i.a. ha-1 de sulfometuron methyl em estádios de desenvolvimento V4 e V8. A produtividade de grãos foi menor apenas para as plantas submetidas às subdoses de sulfometuron methyl, com menos efeito quando a deriva ocorre no estádio de desenvolvimento V8 em comparação ao estádio V4.

Este trabalho busca entender a diversificação craniana dos macacos do Velho Mundo (Catarrhini) integrando duas abordagens para o estudo da evolução de caracteres complexos: a genética quantitativa e a integração morfológica. A investigação tem três objetivos principais: 1) comparar a magnitude e o padrão das relações entre os caracteres cranianos entre todos os Catarrhini; 2) testar a hipótese de que deriva genética é o único agente responsável pela diversificação craniana; 3) explorar as conseqüências evolutivas da associação entre caracteres. De posse de um banco de dados bastante representativo da diversidade dos macacos do Velho Mundo (39 medidas cranianas de cerca de 6.000 crânios de mais de 130 espécies), gerei as matrizes de correlação e de variância/covariância, que resumem as relações entre os caracteres, e comparei-as entre vários grupos. Comparei-as também a expectativas derivadas de modelos teóricos de evolução por deriva genética, além de simular a ação de seleção natural sobre essas matrizes para observar o comportamento evolutivo dos diversos padrões de associação entre caracteres. De maneira geral, o padrão das relações é o mesmo entre todos os Catarrhini, mas a magnitude com que os caracteres estão associados varia bastante. Isso tem conseqüências evolutivas importantíssimas, pois grupos com baixas magnitudes tendem a responder na mesma direção em que a seleção atua (alta flexibilidade evolutiva), enquanto que altas magnitudes estão associadas, independentemente da direção da seleção, a respostas ao longo do eixo de maior variação, que no caso dos Catarrhini corresponde à variação no tamanho (baixa flexibilidade evolutiva). A diversificação inicial do grupo parece ter sido gerada por seleção natural, mas nos níveis de gênero e espécie, deriva genética é o processo predominante; a exceção são os cercopitecíneos, onde há evidência de seleção também nesses níveis. Com base nesses resultados, proponho um modelo que associa a magnitude geral da correlação entre caracteres aos possíveis caminhos evolutivos que uma população pode seguir. Apesar de este trabalho estar empiricamente restrito aos macacos do Velho Mundo, esse modelo é válido para os mamíferos como um todo e pode ser testado em outros grupos, aumentando nossa compreensão de como a associação entre caracteres afeta a evolução dos seres vivos. / This is a study on the cranial diversification of the Catarrhini, a large group of primates that includes all Old World monkeys and apes, bringing together two approaches to investigate the evolution of complex characters: quantitative genetics and morphological integration. It has three main goals: 1) to compare magnitudes and patterns of inter-trait relationships in the skull among catarrhines; 2) to test the null hypothesis that genetic drift is the sole agent responsible for cranial diversification; 3) to explore the evolutionary consequences of inter-trait associations. With a large and representative cranial database of Old World monkeys and apes (39 measurements of around 6,000 skulls from more than 130 species), I generated and compared correlation and variance/covariance matrices, which summarize inter-trait relationships, among several Catarrhini groups. I compared some of those matrices to expectations derived from theoretical models of evolution through genetic drift, and simulated natural selection to observe the evolutionary behavior of each matrix. From a broad perspective, the patterns of relationships are the same among all catarrhines, but the magnitudes are quite variable. This has very important evolutionary consequences, because groups with low overall magnitudes tend to respond in the same direction of selection (high evolutionary flexibility), while higher magnitudes, regardless of the direction of selection, are associated to responses along the axis of highest variation, which in this case corresponds to size variation (low evolutionary flexibility). The initial diversification of catarrhines seems to have been generated by natural selection, but drift probably played a major role at the genus and species level; the exception are the cercopithecines, for which there is evidence for selection also in those levels. Based on these results, I propose a model that links the overall magnitude of inter-trait correlations to the possible evolutionary paths of a given population. This study is empirically restricted to Old World monkeys and apes, but the model has been proved valid to a broader sample of mammals and can be tested for other groups, contributing for our understanding of how complex characters evolve.

O objetivo deste trabalho é avaliar as variações espaço-temporais da dinâmica morfo-sedimentar do sistema praial de Ilha Comprida-SP. Assim, para reconhecer os padrões de onda encontrados na região, resultados do modelo global de ondas WAVEWATCH III foram utilizados. Da mesma forma, levantamentos de campo periódicos foram realizados em cinco diferentes setores com obtenção de dados morfológicos e sedimentares. Simulações com o modelo numérico MIKE21 SW foram realizadas para reconhecer as tendências de transporte e distribuição de força de onda. Os resultados mostraram uma tendência geral de transporte para NE, com maiores valores para a região sul em relação a norte, com dois setores principais de divergência residual: centro-sul e norte. A divergência centro-sul apresenta um aumento e uma migração para sul nos meses mais energéticos, enquanto que a divergência norte apresenta uma manutenção espacial ao longo das estações. A distribuição de força de onda mostrou-se mais elevada nos setores centro-sul e norte, coincidindo com os pontos de divergência. Dados granulométricos mostraram baixa variabilidade temporal, mas possuem diferenças espaciais que refletem a hidrodinâmica local. Os dados volumétricos apresentam maiores valores médios nos setores centro-norte e menores no centro-sul e norte. De modo geral, os resultados mostram que ao longo de uma linha de costa exposta a um mesmo padrão de ondas, sua variabilidade é um reflexo de alterações hidrodinâmicas locais. Tais conclusões são importantes para o conhecimento do estado atual das praias e de sua evolução. / The aim of this study is to evaluate the spatial-time variations on the beach system of Ilha Comprida-SP. Therefore, the wave reanalysis database from the global wave generation model WAVEWATCH III was used to recognize the wave pattern for the region. Furthermore, periodic field works at five different sectors have been conducted to collect morphological and sedimentological data. The numerical model MIKE 21 SW has been applied to propagate waves onshore and recognize the transport tendency and the nearshore wave power distribution. Results show a transport trend to the NE, with the southern sector being larger than the northern sector in magnitude, with two main residual longshore drift divergence spots: in the central-southern and northern regions. Moreover, the central-southern divergence spot become larger and migrated to the south during the most energetic months, while the northern divergence spot kept its position throughout the year. Wave power results show two main areas with higher values that coincide with the observed longshore divergence spots. Sediment data presents low temporal variability, although spatial variations have been found reflecting the hydrodynamic conditions. The volumetric data shows largest values in the central-northern sector, being smaller in the central-southern and northern regions. In summary, the findings show that along this wide open stretch of coastline, exposed to the same offshore wave regime, its variability is a result of local hydrodynamic conditions. These results help in further understanding the island\'s long term evolution and current state of its beaches.

O composto GN-04 foi utilizado no presente trabalho, tendo por objetivo: a) avaliar o efeito do processo dispersivo em subpopulações de milho com diferentes tamanhos efetivos, e submetidas previamente à seleção divergente para alta e baixa produção; b) avaliar a capacidade de combinação e o potencial heterótico dessas subpopulações e c) estimar componentes da variância genética no composto GN-04, usando o esquema de cruzamento em cadeia. Um lote do referido composto foi plantado visando à obtenção de três diferentes progênies: S1, por autofecundação; irmãos germanos, cruzamento planta a planta e cruzamentos em cadeia, seguindo o esquema 1 x 2, 2 x 3 e 3 x 1. Foram obtidas 77 progênies S1, 75 de irmãos germanos e 72 cadeias, contituidas de três cruzamentos. As progênies de cada tipo foram avaliadas em ensaios distintos, em dois locais, de acordo com o delineamento em blocos casualizados, com três repetições. Os caracteres avaliados foram: altura da planta (AP), altura da espiga (AE), comprimento da espiga (CE), diâmetro da espiga (DE), peso de espigas (PE) e prolificidade (PR). Com base nos valores médios do caráter PE, foram selecionadas 10 progênies de irmãos germanos,10 de S1 e 10 cadeias, sendo cada grupo de materiais constituído das cinco progênies ou cadeias mais produtivas, e das cinco menos produtivas. A recombinação das progênies selecionadas ou cadeias ensejou a obtenção de subpopulações com diferentes tamanhos efetivos: N1, N2 e N3, que correspondem a taxas de endogamia de 50%, 25% e 17%, para S1’s, irmãos germanos e cadeias, respectivamente. Essas subpopulações foram avaliadas posteriormente com base no seu desempenho per se e em cruzamentos ‘topcrosses’ e dialélicos. Os resultados desta avaliação mostraram que o tamanho reduzido das subpopulações provocou o surgimento de valores médios inferiores ou superiores à média estimada da população original, em conseqüência da deriva genética. Os valores médios de todos os caracteres foram afetados pela redução no tamanho efetivo das subpopulações, em conseqüência da depressão por endogamia, a qual se expressou com mais intensidade no caráter PE: em subpopulações com N3 e N1, no nível de seleção positiva, o efeito da depressão foi de 5,20% e 20,07%, respectivamente, e no nível de seleção negativa, 15,98% e 55,76%, respectivamente. A seleção divergente entre subpopulações resultou em diferenciação genética suficiente para expressar heterose em cruzamentos entre subpopulações com o mesmo tamanho efetivo. As estimativas da variância aditiva do composto GN-04, com base na análise da variância conjunta dos experimentos com progênies de cruzamentos em cadeia, para os caracteres PE, CE, DE, PR, AP e AE foram, respectivamente: 128,00 (g/pl)², 1,490 (cm/esp.)², 0,1704 (cm/esp.)², 0,0200 (esp./pl)², 256,55 (cm/pl)² e 136,80 (cm/pl)². Estimativas negativas da variância aditiva foram obtidas para os caracteres CE, DE, AP e AE; para os caracteres PE e PR as estimativas foram 185,00 (g/pl)² e 0,0188 (esp./pl)², respectivamente. O delineamento de cruzamentos em cadeia de tamanho 3, tendo em vista a facilidade de execução, pode ser uma alternativa a delineamentos de execução mais complexa para estimar componentes da variância. / The composite GN-04 was used in the present work with the objetives: a) evaluation of the effect of the dispersive process in subpopulations of maize with different effective sizes, and submitted to divergent selection for higth and low yield; b) evaluation of the combining ability and heterotic potential of the subpopulations; and c) to estimate components of the genetic variance in the composite GN-04, using chain crosses. A pollination block of the base population was planted for the development of three different progenies: S1 progenies obtained by selfing, full-sib progenies obtained by plant-to-plant crosses, and chain crosses following the scheme: 1 x 2, 2 x 3 e 3 x 1. The numbers of progenies were: 77 S1´s, 75 full-sibs and 72 chains represented by three crosses in each chain. The progenies from each type were evaluated in different experiments in two locations following the randomized complete block design with three replications. The following traits were analysed: plant height (AP), ear height (AE), ear length (CE), ear diameter (DE), ear weight (PE) and prolificacy (PR). Observed means of PE were used for selection of 10 full-sib progenies, 10 S1´s progenies and 10 complete chains, each group being represented by five high yielding and five low yielding progenies or chains. The recombination of the selected progenies or chains, individually, led to the development of subpopulations with different effective sizes: N1, N2 and N3, corresponding to inbreeding rates of 50%, 25% and 17% for S1´s, full-sibs and chain crosses, respectively. The subpopulations were later evaluated based on their performance per se, in the topcrosses and in diallel crosses. Results showed that the reduced size of the subpopulations led to changes in the means of the evaluated traits, which were smaller or higher than the original means as a consequence of genetic drift. On the average, all traits showed a decrease in the mean as a consequence of inbreeding depression. The most depressive trait was PE and the depression effects in subpopulations with N3 and N1 were 7.17% and 22.94% in the positive selection and 16.85% and 55.91% in the negative selection. The divergent selection among subpopulations led to a genetic differentiation sufficient for the expression of heterosis in crosses between subpopulations in the same level of effective size. The estimates of the additive genetic variance in the composite GN-04 were obtained from the analysis of chain crosses. The estimates for the traits PE, CE, DE, PR, AP and AE were, respectively: 128.00 (g/pl)², 1.490 (cm/esp.)², 0.1704 (cm/esp.)², 0.0200 (esp./pl)², 256.55 (cm/pl)² and 136.80 (cm/pl)². Negative estimates of the dominance variance were obtained for the traits CE, DE, AP and AE; for the traits PE and PR, the estimates were 185.00 (g/pl)² and 0.0188 (esp./pl)², respectively. The mating design based on chain crosses with size 3 was considered feasible and easy to be used and can be a reliable alternative as compared with other more complex designs.

In the software development life cycle, changes are inevitable. Designing the architecture of the software and writing the source code does not end the software life cycle. The software system evolves as changes in the environment and requirements are incorporated in the system. If these changes are not managed properly, the architecture of the software deteriorates and leads to architecture erosion. This study is an effort to address the problem of architecture erosion and to keep the software architecture live and up to date. In this study we propose a method to minimize or avoid architecture erosion. This method provides a systematic way to design the architecture of the software system. It is based on the concept of architecture vision that is the ideal representation of the architecture of the software in hand. This method treats architecture document and architecture design document as two completely different documents. Moreover this study presents an effort to establish a relationship between the implemented architecture of the system, the architecture design document and an architecture vision. The dynamic validation of the devised method shows that this method is suitable for medium to large scale projects that have several releases. All the activities of this method revolve around the efforts to keep the architecture of the software aligned with an architecture vision. The iterative nature of the method and synchronization of the implemented architecture with the architecture vision helps to detect and reduce architecture erosion. This method does not conflict with or replaces any development or management process or method (like RUP, agile, water fall etc), but exists parallel to them. / 0739 229 759 - Aamir 0737 351 062 - Mustafa

Föreliggande studie har undersökt flera gränssnitt i en helikoptersimulator som visar horisontell och vertikal drift; en taktil, en visuell och en bimodal display. Syftet med studien var att undersöka om det är med fördel att använda displayer som visar på drift horisontellt och vertikalt. Dessutom var syftet att undersöka om taktil display leder till lika bra prestation som med en visuell display, eller kommer kunna användas som komplement till den visuella i en bimodal kombination. I studien deltog 12 försöksdeltagare som fick i uppgift att i en helikoptersimulator försöka stå stilla i luften (s.k. hovra) på en höjd på 8000 fot med hjälp av varje driftdisplay under två minuter. Förutom de tre driftdisplayerna skulle även försöksdeltagarna hovra utan några hjälpmedel som visar på drift. Resultatet visade att det är med fördel att använda alla tre driftdisplayer för att minska drift horisontellt. Resultatet visade dessutom att det är med fördel att använda en visuell display och en bimodal display för att minska drift vertikalt. Slutsatsen av studien är således att det är med fördel att använda driftdisplayer för att minska på drift. Den taktila displayen fungerar inte lika bra som den visuella i undvikandet av drift, men går att användas som komplement i en bimodal kombination.

Billions of dollars of loss are caused every year by fraudulent credit card transactions. The design of efficient fraud detection algorithms is key for reducing these losses, and more and more algorithms rely on advanced machine learning techniques to assist fraud investigators. The design of fraud detection algorithms is however particularly challenging due to the non-stationary distribution of the data, the highly unbalanced classes distributions and the availability of few transactions labeled by fraud investigators. At the same time public data are scarcely available for confidentiality issues, leaving unanswered many questions about what is the best strategy. In this thesis we aim to provide some answers by focusing on crucial issues such as: i) why and how undersampling is useful in the presence of class imbalance (i.e. frauds are a small percentage of the transactions), ii) how to deal with unbalanced and evolving data streams (non-stationarity due to fraud evolution and change of spending behavior), iii) how to assess performances in a way which is relevant for detection and iv) how to use feedbacks provided by investigators on the fraud alerts generated. Finally, we design and assess a prototype of a Fraud Detection System able to meet real-world working conditions and that is able to integrate investigators’ feedback to generate accurate alerts. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

Det här arbetet bygger på en litteraturstudie från föregående arbeten som handlar om biobränslet HVO och dess funktion vid drift av flera olika typer av dieselmotorer. Författarna till arbetet har tittat på de för- och nackdelar som föreligger vid drift på HVO och gjort jämförelser med diesel och HFO vilka dominerar inom sjöfarten idag. Arbetet har gjorts med fokus på teknisk drift samt miljöaspekter vid användning och vi har bortsett från annat som tillgång, framställning och distribution. Vi har även tagit upp kommande utsläppskrav inom sjöfarten och diskuterat vad som kan behöva göras för att klara av dessa. Skillnader i utsläpp vad gäller koldioxid, kväveoxider, svaveloxider och partiklar behandlas tillsammans med driftsaspekter som prestanda, bränsleförbrukning och lagringsstabilitet. Sett till sin helhet har resultaten visat sig vara goda för HVO vid jämförelser mot diesel och HFO, särskilt ur ett miljöperspektiv. Det största dilemmat i dagsläget är huruvida man ska kunna tillgodose de enorma behoven som råder på bränsle, då råvarutillgången är begränsad och framställningsmetoderna kostsamma. / This work is based on a literature study from previous work regarding the biofuel HVO and its function in the operation of several different types of diesel engines. The authors of this work have looked at the pros and cons of operating at HVO and made comparisons with diesel and HFO which dominate in shipping today. The work has been done with a focus on technical operations and environmental aspects in use and we have not taken availability, production and distribution into consideration. We have also raised future emission requirements in shipping and discussed what may be needed to cope with these. Differences in carbon dioxide, nitrogen oxides, sulfur oxides and particulate emissions are treated together with operational aspects such as performance, fuel consumption and storage stability. In summary, the results have proven to be in favor for HVO in comparison to diesel and HFO, especially from an environmental perspective. The biggest dilemma at present is whether to meet the enormous needs of fuel, as raw materials are limited and the production methods costly.

Wiener process-a random process with continuous time-plays an important role in mathematics, physics or economy. It is often good to know whether it contains any deterministic part, e.g. drift or scale. However, it is nearly impossible either observe the whole trajectory of the process or preserve its full history. This thesis deals with a statistical analysis based on partial observations, namely passage times through some given barriers. We propose several statistical methods for testing hypotheses about drift or scale using these observations. As supporting methods, we consider the maximum likelihood theory, non-parametric test against a trend, and binomial test. For testing the value of scale in the model with no drift and constant scale we recommend maximum likelihood theory. We derive the estimate and related tests in the case of observing only three barriers. The simulation study suggested observing more barriers for testing monotony of scale in a model with linear drift, or testing monotone and convex/concave drift in a model with constant scale. 1

Une plante peut être totalement protégée d'un agent pathogène grâce à un gène majeur de résistance, mais ce dernier peut être rapidement contourné suite à l'apparition et à la propagation de variants pathogènes adaptés. Cette thèse s'intéresse aux mécanismes évolutifs permettant le ralentissement de ce contournement chez les virus de plantes en agissant sur deux forces évolutives majeures, la dérive génétique et la sélection, depuis le niveau de l'hôte jusqu'à celui de la parcelle. D'abord, un modèle épidémiologique stochastique de type SI au niveau d'une parcelle agricole a montré que la dérive génétique pouvait être particulièrement bénéfique au rendement agricole lorsque l'adaptation du virus au gène majeur induit un coût de fitness intermédiaire dans les plantes sensibles. Ensuite, la conception et la validation d'un modèle basé sur des équations déterministes de Lotka-Volterra et des processus stochastiques Dirichlet-multinomiaux a permis de distinguer les effets de la dérive génétique et ceux de la sélection sur des données temporelles de compétition intra-plante entre variants viraux, et de mettre en évidence le contrôle génétique de ces effets par les plantes. Enfin, une analyse de la corrélation entre ces estimations des intensités de dérive génétique et de sélection et une estimation expérimentale de la durabilité d'un gène majeur a montré qu'une forte dérive génétique lors des stades précoces de l'infection augmentait la durabilité du gène majeur. Ces résultats ouvrent des perspectives pour une gestion plus durable de la résistance des plantes, par la sélection de variétés de plantes induisant une forte dérive génétique sur les populations d'agents pathogènes / Plants can be fully protected from their pathogens when they carry major resistance genes, but the efficiency of these genes is limited by the emergence and spread of adapted, resistance-breaking pathogen variants. This thesis studies how evolutionary forces imposed by the plants on pathogen populations may increase the durability of major resistance genes. Using plant viruses as a biological model, this thesis investigates the effect of genetic drift and selection, from the within-host to the host population level. Firstly, a stochastic epidemiological SI model at the field level showed that genetic drift could be particularly beneficial for crop yield when the fitness cost associated with virus adaptation to resistance was intermediate in susceptible plants. Then, the design and validation of a mechanistic-statistical model based on deterministic Lotka-Volterra equations and stochastic Dirichlet-multinomial processes allowed to disentangle the effects of genetic drift from those of selection on temporal data of within-host competition between virus variants. The intensities of genetic drift and selection acting on virus populations were shown to be controlled genetically by the hosts. Finally, a correlation analysis between these estimations of genetic drift and selection intensities and an experimental estimation of the durability of a major resistance gene showed that strong genetic drift during the early stages of plant infection led to an increase in resistance durability. These results open new perspectives for more durable management of plant resistance, by breeding plant varieties inducing strong genetic drift on pathogen populations