京都大学 / 新制・課程博士 / 博士(理学) / 甲第24310号 / 理博第4880号 / 新制||理||1698(附属図書館) / 京都大学大学院理学研究科生物科学専攻 / (主査)教授 曽田 貞滋, 准教授 渡辺 勝敏, 教授 中務 真人 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

Deep Learning (DL) has been broadly applied to solve big data problems in biomedical fields, which is most successful in image processing. Recently, many DL methods have been applied to analyze genomic studies. However, genomic data usually has too small a sample size to fit a complex network. They do not have common structural patterns like images to utilize pre-trained networks or take advantage of convolution layers. The concern of overusing DL methods motivates us to evaluate DL methods' performance versus popular non-deep Machine Learning (ML) methods for analyzing genomic data with a wide range of sample sizes. In this paper, we conduct a benchmark study using the UK Biobank data and its many random subsets with different sample sizes. The original UK Biobank data has about 500k participants. Each patient has comprehensive patient characteristics, disease histories, and genomic information, i.e., the genotypes of millions of Single-Nucleotide Polymorphism (SNPs). We are interested in predicting the risk of three lung diseases: asthma, COPD, and lung cancer. There are 205,238 participants have recorded disease outcomes for these three diseases. Five prediction models are investigated in this benchmark study, including three non-deep machine learning methods (Elastic Net, XGBoost, and SVM) and two deep learning methods (DNN and LSTM). Besides the most popular performance metrics, such as the F1-score, we promote the hit curve, a visual tool to describe the performance of predicting rare events. We discovered that DL methods frequently fail to outperform non-deep ML in analyzing genomic data, even in large datasets with over 200k samples. The experiment results suggest not overusing DL methods in genomic studies, even with biobank-level sample sizes. The performance differences between DL and non-deep ML decrease as the sample size of data increases. This suggests when the sample size of data is significant, further increasing sample sizes leads to more performance gain in DL methods. Hence, DL methods could be better if we analyze genomic data bigger than this study. / Graduate

Globally disseminated bacterial pathogens frequently cause epidemics that are of major importance in public health. Of particular significance is the capacity for some of these bacteria to switch into a new environment leading to the emergence of pathogenic clones. Understanding the evolution and epidemiology of such pathogens is essential for designing rational ways for prevention, diagnosis and treatment of the diseases they cause. Whole-genome sequencing of multiple isolates facilitating comparative genomics and phylogenomic analyses provides high-resolution insights, which are revolutionizing our understanding of infectious diseases. In this thesis, a range of population genomic analyses are employed to study the molecular mechanisms and the evolutionary dynamics of bacterial pathogen niche adaptation, specifically between humans, animals and the environment. A large-scale population genomic approach was used to provide a global perspective of the host-switching events that have defined the evolution of Staphylococcus aureus in the context of its host-species. To investigate the genetic basis of host-adaptation, we performed genome-wide association analysis, revealing an array of accessory genes linked to S. aureus host-specificity. In addition, positive selection analysis identified biological pathways encoded in the core genome that are under diversifying selection in different host-species, suggesting a role in host-adaptation. These findings provide a high-resolution view of the evolutionary landscape of a model multi-host pathogen and its capacity to undergo changes in host ecology by genetic adaptation. To further explore S. aureus host-adaptive evolution, we examined the population dynamics of this pathogen after a simulated host-switch event. S. aureus strains of human origin were used to infect the mammary glands of sheep, and bacteria were passaged in multiple animals to simulate onward transmission events. Comparative genomics of passaged isolates allowed us to characterize the genetic changes acquired during the early stages of evolution in a novel host-species. Co-infection experiments using progenitor and passaged strains indicated that accumulated mutations contributed to enhanced fitness, indicating adaptation. Within-host population genomic analysis revealed the existence of population bottlenecks associated with transmission and establishment of infection in new hosts. Computational simulations of evolving genomes under regular bottlenecks supported that the fitness gain of beneficial mutations is high enough to overcome genetic drift and sweep through the population. Overall, these data provide new information relating to the critical early events associated with adaptation to novel host-species. Finally, population genomics was used to study the total diversity of Legionella longbeachae from patient and environmental sources and to investigate the epidemiology of a L. longbeachae outbreak in Scotland. We analysed the genomes of isolates from a cluster of legionellosis cases linked to commercial growing media in Scotland and of non-outbreak-associated strains from this and other countries. Extensive genetic diversity across the L. longbeachae species was identified, associated with intraspecies and interspecies gene flow, and a wide geographic distribution of closely related genotypes. Of note, a highly diverse pool of L. longbeachae genotypes within compost samples that precluded the genetic establishment of an infection source was observed. These data represent a view of the genomic diversity of this pathogen that will inform strategies for investigating future outbreaks. Overall, our findings demonstrate the application of population genomics to understand the molecular mechanisms and the evolutionary dynamics of bacterial adaptation to different ecological niches, and provide new insights relevant to other major bacterial pathogens with the capacity to spread between environments.

Le sujet de ma thèse était de décrire sur le plan moléculaire le processus d'embryogenèse somatique chez la vigne. Pour cela, les étapes-clés d'entrée et de sortie du cycle d'embryogenèse secondaire ont été caractérisées par l'analyse de l'expression de quelques gènes impliqués dans le développement ou la défense, en particulier les gènes PR10.Grâce à l'exploitation de la séquence complète du génome de Vitis vinifèra disponible sur le site du Genoscope, j'ai pu caractériser exhaustivement la famille multigènique des PR10. Celle-ci est composée de 17 séquences disposées en tandem et formant un cluster compact sur le chromosome 5, dont 3 pseudogènes et au moins 13 séquences transcrites. L'expression de 10 de ces gènes a d'abord été analysée par RT-PCR semi-quantitative dans différents organes de la plante et dans des tissus traités au 2,4-D. Elle suggère une diversification fonctionnelle marquée. De plus, le niveau d'expression de plusieurs gènes PR10 est élevé dans les cals embryogènes, suggérant qu'ils pourraient jouer un rôle lors de l'embryogenèse somatique. L'étude de l'expression des gènes PR10 par RT-PCR quantitative en temps réel dans différents tissus ayant montré une capacité embryogénique variable lorsqu'ils sont soumis à un traitement par le 2,4-D met en évidence que le niveau d'expression varie entre les gènes et selon les tissus. L'expression de certains gènes est fortement induite par le 2,4-D dans les tissus à capacité embryogénique et seulement faiblement dans les tissus ne donnant jamais d'embryons somatiques, ce qui suggère fortement que ceux-ci pourraient être des marqueurs de la capacité embryogénique chez la vigne. / The objective of my work was to analyse the somatic embryogenesis process of Vitis vinifera at a molecular scale. Thus, the expression of genes implied in development or defence, especially PR10 genes, was monitored during the key-steps of entrance and exit of secondary somatic embryogenesis. The complete sequence of the Vitis vinifera genome available on the Genoscope website allowed the exhaustive characterization of the PR10 multigene family, which is constituted by 17 sequences localised on a tandem array on the chromosome 5. Among these 17 sequences, 3 are pseudogenesand at !east 13 are transcribed sequences. The expression of 10 PR10 genes was first monitored in various grapevine tissues and in tissues after 2,4-D treatment using semi-quantitative RT-PCR. The results suggest a strong functional diversification. Moreover, the expression of several PR10 genes is high in embryogenic calli, suggesting that these genes could intervene in somatic embryogenesis. The expression of PR10 genes was also monitored in tissues showing different somatic embryogenic capabilities under 2,4-D treatment using quantitative RT-PCR. The results show that regulation of PR10 genes is dependent of the gene and tissue considered. Moreover, the expression of some genes is highly induced by 2,4-D treatment in tissues having embryogenic capability, white it is only weakly induced in tissues having no embryogenic capability, suggesting that these gene could be markers of embryogenic capability in grapevine.

No description available.

L'agricoltura moderna sta affrontando sfide come la perdita di fertilità del suolo, la variabilità climatica e gli attacchi di agenti patogeni in continuo aumento. Le pratiche agricole si stanno evolvendo verso sistemi sostenibili e rispettosi dell'ambiente. L'uso di biostimolanti (PBS, plant biostimulant) è una soluzione innovativa per affrontare le sfide di un’agricoltura sostenibile che garantisce un assorbimento ottimale dei nutrienti, una resa delle colture e tolleranza agli stress abiotici. In particolare, tra i diversi tipi di biostimolanti presenti sul mercato, i rizobatteri, classificati come Plant Growth Promoting Rhizobacteria (PGPR), offrono un nuovo approccio per promuovere la crescita delle piante, la mitigazione degli stress e l’aumento della resa colturale. Pertanto i PGPR sono considerati come una sorta di "probiotici" vegetali, poiché contribuiscono in modo efficiente alla nutrizione e all'immunità delle piante. L'obiettivo principale di questa tesi è isolare e identificare batteri presenti nella rizosfera di pomodoro (Solanum lycopersicum L.) che mostrano proprietà PBS, nonché valutare i meccanismi coinvolti nell'azione di promozione della crescita delle piante (Capitolo 2) e la genetica alla base di questi meccanismi (Capitolo 3 e 4). Infatti, una profonda comprensione dei meccanismi d’azione dei PGPR potrebbe colmare la mancanza di coerenza del dato di efficacia tra gli studi di laboratorio e gli studi in campo e stimolare la ricerca per la produzione e la commercializzazione di nuovi prodotti biostimolanti microbici. / Modern agriculture faces challenges such as loss of soil fertility, fluctuating climatic factors and increasing pathogen and pest attacks. Agricultural practices have been evolving towards organic, sustainable and environmentally friendly systems. The use of natural plant biostimulants (PBS) is an innovative solution to address the challenges in sustainable agriculture, to ensure optimal nutrient uptake, crop yield, quality and tolerance to abiotic stress. In particular, among different types of biostimulants present on the market, plant growth promoting rhizobacteria (PGPR) offer a novel approach for promoting plant growth, mitigate stress and increase crop yield. Hence, PGPR inoculants are now considered as a kind of plant ‘probiotics’, since they efficiently contribute to plant nutrition and immunity. The main goal of this thesis was to isolate and identify bacteria symbionts of tomato (Solanum lycopersicum L.) rhizosphere, which showed PBS properties and evaluate mechanism involved in the action of PGPR (Chapter 2), underlying genetics and physiological pathways (Chapter 3 and 4). Indeed, a deeply understanding of the mechanisms of plant growth promotion, could fulfill the lack of consistency between lab, greenhouse and field studies, and support commercialization of novel plant biostimulant products.

La recherche est imbriquée dans les soins cliniques en oncologie, de sorte que la dualité de rôles (clinicien, investigateur) est une réalité connue des oncologues. Vu l'importance de la génomique pour guider les soins en oncologie pédiatrique, la plupart des patients sont enrôlés dans des études génomiques au Québec. En pratique, l'oncologue, par sa présence aux réunions d'oncogénomique, participe avec les chercheurs à l'interprétation des résultats génétiques ainsi qu'aux décisions de divulgation au patient. Ce projet vise à identifier et caractériser les enjeux éthico-légaux soulevés par la dualité de rôles de l'oncologue durant ce partage de résultats génétiques en amont du patient. Un questionnaire incluant six vignettes narratives a été diffusé électroniquement à tous les oncologues pédiatres du Québec. Des entrevues semi-dirigées ont été effectuées avec un sous-groupe de participants. Vingt-huit oncologues ont complété le questionnaire et cinq oncologues ont participés aux entrevues. Les niveaux de confort des oncologues étaient influencés par le type de résultat génétique, par le contenu discuté lors du consentement et par le rôle de l'oncologue envers le patient (oncologue traitant ou non). Le fait d'être informé d'un résultat génétique de recherche suffisait pour déclencher un sentiment de responsabilité par rapport à celui-ci. La dualité de rôles est incontournable et donne à l’oncologue un accès privilégié à l'information génétique, au-delà de ce à quoi aura accès le patient. Les responsabilités et les devoirs de l’oncologue dans le cadre de la relation thérapeutique sont au centre des enjeux éthiques et légaux soulevés par la dualité de rôles. / Research is heavily integrated in oncology care, with oncologists well acquainted with the concept of duality of roles (doctor, researcher). Given the importance of genomics in pediatric oncology, most patients are offered genomic sequencing via research initiatives in Quebec. Practical experience reveals that oncologists, by attending molecular tumour board meetings, participate with the research team in the interpretation of genetic research results and decisions regarding disclosure to patients. This project aims to identify and characterise the ethical and legal issues related to the duality of roles of oncologists during this process of genetic information sharing, prior to informing the patient. A questionnaire including six narrative case vignettes was electronically distributed to all pediatric oncologists in Quebec. A semi-structured interview was then conducted with a sub-group of participants. Twenty-eight oncologists completed the questionnaire and five oncologists participated in the interviews. Oncologists' comfort levels were influenced by the type of genetic result, by the content of prior consent discussions and by their specific role regarding the patient (treating oncologist or not). The state of becoming aware of a result was sufficient to trigger a feeling of responsibility regarding that genetic research result. Duality of roles is inevitable and provides the oncologist with privileged access to genetic information, above what will be accessible to the patient. Responsibilities and duties of the oncologist in the setting of a therapeutic relationship are central to the ethical and legal issues raised by the duality of roles.

L'augmentation et la dissémination de la résistance aux antibiotiques chez les bactéries à gram-negatif, particulièrement les Entérobactéries, les bactéries du genre Pseudomonas et Acinetobacter, représentent un problème majeur de santé publique au niveau mondial. Les infections nosocomiales causées par les bactéries multi-résistantes (BMR) ont conduit non seulement à une augmentation de la mortalité, de la morbidité, et du coût de traitement, mais aussi continuent de mettre en danger la vie des patients surtout immunodéprimés en milieu hospitalier. Bien entendu, l'utilisation abusive et non contrôlée des antibiotiques a grandement contribué à la large diffusion des déterminants de la résistance; cependant, des études récentes ont démontré que ces déterminants de la résistance pouvaient émerger à partir de sources anciennes et/ou environnementales. Ainsi, face à cette préoccupation mondiale, plusieurs études ont été rapportées avec des recommandations importantes de conduire des études épidémiologiques, moléculaires, et génomiques afin de contrôler la diffusion et l'augmentation de la résistance aux antibiotiques. De plus, durant ces 10 dernières années, nous avons assisté à l'emergence et au développement de nouvelles technologies de séquençage à haut débit coïncidant avec une augmentation exponentielle du nombre de genomes bactériens séquencés. / The increase and spread of multidrug-resistant (MDR) gram-negative bacteria especially Enterobacteriaceae, Pseudomonas, and Acinetobacter (E.P.A) species have become a major concern worldwide. The hospital-acquired infections caused by MDR bacteria have led not only to an increase in mortality, morbidity, and cost of treatment, but also continue to endanger the life of patients, especially those immunocompromised. Although the frequent misuse of antibiotic drug has greatly contributed to worldwide dissemination and resistance to antibiotics; recent studies have shown that these resistance determinants could emerge from ancient or environmental sources. Front of this worldwide concern, several studies have been reported with significant recommendations to conduct molecular epidemiology, and genomic studies, in order to control the increase and the dissemination of the antibiotic resistance. Moreover, during these last 10 years, we are witnessing the emergence and development of new technologies of high throughput sequencing and coinciding with an exponential increase of number of bacterial genomes sequenced today. Therefore, it is in this context that the project of this thesis was conducted with three essential objectives: (i) the genome sequencing of clinical MDR bacteria, the analysis and the identification of the mechanisms and the genetic determinants of antimicrobial resistance (ii) the achievement of molecular epidemiology studies from clinical MDR bacteria responsible of outbreak (iii) the development and implementation of molecular tools for monitoring and diagnosis of potential MDR bacteria.

No description available.

La détection, la surveillance et la diffusion de la résistance des bactéries aux antibiotiques est un enjeu majeur au niveau mondial depuis la découverte et la diffusion de bactéries multi résistantes, en particulier la résistance aux carbapénèmes, spécifiquement chez les Entérobactéries et les bactéries du genre Pseudomonas et Acinetobacter. L’émergence et la dissémination des pathogènes Gram- résistants aux carbapénèmes est un contributeur significatif de la morbidité et la mortalité du patient. Malgré les efforts radicaux dans le contrôle de l’infection et les améliorations dans le diagnostique moléculaire, les bacilles Gram- résistants aux carbapénèmes demeurent une formidable menace vue que quelques agents antimicrobiens sont actifs et très peu devraient être disponibles dans le futur proche.L’origine et la source des gènes de résistance dans le monde sont mal connues et des travaux récents suggèrent que les animaux domestiques et sauvages, l’environnement mais également le tube digestif des mammifères et des humains pourraient représenter un réservoir et une source importante de gènes de résistance susceptibles d’être transmissibles à l’homme. C’est dans cette optique que ce projet de thèse s’articule avec comme objectifs : (i) la réalisation d’études épidémiologiques moléculaires d’isolats cliniques et animales résistants aux carbapénèmes isolés dans le basin Méditerranéen et la caractérisation des supports moléculaires de cette résistance ; (ii) la description de nouveaux mécanismes de résistance a l’imipenème ; et enfin (iii) le séquençage de génomes d’isolats cliniques résistants aux carbapénèmes et l’analyse de ces derniers. / The detection, monitoring and dissemination of bacterial resistance to antibiotics are a major issue worldwide since the discovery and spread of multi-resistant bacteria, in particular resistance to carbapenems, specifically among Enterobacteriaceae and bacteria of the genus Pseudomonas and Acinetobacter.The emergence and dissemination of carbapenem-resistant Gram-negative pathogens is a significant contributor to patient morbidity and mortality. Despite radical efforts in infection control and improvements in molecular diagnostics, carbapenem-resistant Gram-negative bacilli remain a formidable threat as few antimicrobial agents are reliably active and very little is expected to be available in the near future.The origin and source of resistance genes in the world are not well known and recent works suggest that domestic and wild animals, the environment (soil, water, rivers ..) but also the digestive tract of mammals and humans could represent a reservoir and an important source of resistance genes that may be transmissible to humans.It is in this context that this thesis project articulates with the following objectives: (i) The achievement of molecular epidemiological studies on carbapenem-resistant clinical and animal isolates collected from countries in the Mediterranean basin (Lebanon, Libya, France) and the characterization of the genetic determinants of this resistance; (ii) the description of new resistance mechanisms to imipenem; and finally (iii) The genome sequencing of clinical isolates resistant to carbapenems, the analysis of these genomes and the identification of mechanisms and genetic supports of the resistance to carbapenems and other antibiotics.