Biobanks pose unique challenges to legal and bioethical frameworks, and raise many as yet unanswered questions, including how these collections of biological samples and information should be governed and for whose benefit. While some commentators have suggested that biobanks should be regulated through specific legislation, I focus on exploring alternative models of governance. I examine, in particular, the interrelationship between benefit-sharing and public engagement, arguing that public engagement is a benefit in itself, valuable both in its own right and as an essential component of good governance, and critically examine proposals for more direct 'representative‘ forms of participant involvement and 'power-sharing‘ arrangements in the biobanking context. Central to my arguments is the concept of the "common heritage", which has been invoked by UNESCO and HUGO in relation to the human genome. From its early beginnings in the law of the sea, this concept has been linked to notions of solidarity, reciprocity and equitable access and sharing. Applied in the context of biobanks, the "common heritage" highlights the value of genetic collections and research for the benefit of present and future generations. Viewed as a third generation human right, the "common heritage" also links to notions of citizenship, civic involvement in policy processes and, ultimately, to participatory or deliberative democracy. From this, I suggest that robust biobank governance mechanisms require not only effective benefit-sharing arrangements but that these must necessarily involve provision for effective public engagement. Drawing on democratic and business management theory, I argue for a 'stakeholder' model of governance. This model draws its basic ideology from communitarian philosophy and regards any organisation (whether it be a corporation or a charity) as a 'social entity', accountable to a broad range of stakeholders. It is my contention that a stakeholder model is the most appropriate model of governance for large-scale population biobanks, such as UK Biobank, which are designed for public benefit, to enhance the health of all, including future generations. In sum, it is a model through which the common interest vested in biobank research might materialise.

Epidemiologic studies have adapted to the genomics era by forming large international consortia to overcome issues of large data volume and small sample size. Whereas both cohort and well-conducted case-control studies can inform disease risk from genetic susceptibility, cohort studies offer the additional advantages of assessing lifestyle and environmental exposure-disease time sequences often over a life course. Consortium involvement poses several logistical and ethical issues to investigators, some of which are unique to cohort studies, including the challenge to harmonize prospectively collected lifestyle and environmental exposures validly across individual studies. An open forum to discuss the opportunities and challenges of large-scale cohorts and their consortia was held in June 2009 in Banff, Canada, and is summarized in this report.

Submitted by Oliveira Santos Dilzaná (dilznana@yahoo.com.br) on 2016-04-04T14:54:45Z No. of bitstreams: 1 Tese de Rosanita Ferreira e Baptista.pdf: 2893108 bytes, checksum: 82b1fb069c85249aab41ed18da8212cb (MD5) / Approved for entry into archive by Ana Portela (anapoli@ufba.br) on 2016-04-06T14:26:35Z (GMT) No. of bitstreams: 1 Tese de Rosanita Ferreira e Baptista.pdf: 2893108 bytes, checksum: 82b1fb069c85249aab41ed18da8212cb (MD5) / Made available in DSpace on 2016-04-06T14:26:35Z (GMT). No. of bitstreams: 1 Tese de Rosanita Ferreira e Baptista.pdf: 2893108 bytes, checksum: 82b1fb069c85249aab41ed18da8212cb (MD5) / As biociências, ao tomarem o corpo humano como objeto de escrutínio e experimentação, mobilizam questões e atores que não são apenas científicos e técnicos, mas também políticos, legais e éticos. A imbricação dessas esferas parece tornar problemáticas as perspectivas que se fundamentam em dualidades, como natureza x cultura, fato x valor. Esse é o caso dos biorrepositórios e biobancos com finalidades de pesquisa, cujas práticas de manusear e colecionar material biológico humano e informações associadas para uso em pesquisas, ao tempo em que gera expectativas para o desenvolvimento da saúde e da medicina, também trazem incertezas e controvérsias sobre a natureza e destino das coleções, bem como sobre as consequências dos experimentos. É um contexto em que as práticas científicas engendram entidades que não se enquadram confortavelmente nas ontologias modernas e, mais especificamente, tornam ambíguas as fronteiras do “humano”, demandando novas regulamentações e ordenações. Assim, sob o referencial da Teoria Ator-Rede, esta tese propõe explorar as controvérsias que tomaram forma nos eventos de construção da regulamentação dos biobancos e biorrepositórios no Brasil, no período de 2009 a 2011, sob condução do Ministério da Saúde (MS) e do Conselho Nacional de Saúde (CNS) e que resultaram na publicação das Diretrizes para Análise Ética de Projetos de Pesquisas que Envolvam Armazenamento de Material Biológico Humano ou Uso de Material Armazenado em Pesquisas (Resolução CNS Nº 441/11) e nas Diretrizes Nacionais para Biorrepositórios e Biobancos com Finalidade de Pesquisa (Portaria do MS nº 2.201/ 11). Dois objetivos principais norteiam a investigação: mapear as controvérsias que envolvem a temática do armazenamento e uso de informações e material biológico humano em pesquisa e apreender a tessitura híbrida que conformaram normas e padrões para estas práticas tecnocientíficas. By adopting the human body as an object of scrutiny and experimentation, the biosciences has mobilized issues and actors who are not only scientific and technical, but also political, legal and ethical. The overlapping of these spheres seems to interfere with prospects that are based on dualities, such as nature vs. culture, fact vs. value. This is the case with regards to biobanks and biorepositories that are intended for research, where the practices of collecting and handling human biological material and associated information for use in research. This also creates expectations for the development of health and medicine, they too bring uncertainties and controversies about the nature and destination of the collections, as well as on the consequences of the experiments. It is a context in which scientific practices create entities that do not fit comfortably in modern ontologies, and more specifically, they make unclear the boundaries of the "human", demanding new regulations and ordinances. Therefore, under the Actor- Network Theory framework, this thesis intends to explore the controversies that took shape in the conception event of the regulation of biobanks and biorrepositórios in Brazil, in the 2009-2011 period, under the guidance of the Ministry of Health (MOH) and the National Health Council (CNS), which resulted in the publication of the Guidelines for ethical analyzing of research projects that involve material storage or utilization of stored materials in prior researches (Ordinance No. 2201) and the Brazilian Resolution on the storage and use of human biological material in research projects (Resolution 441/11). There are two main objectives guiding the research: to map the controversies surrounding the issue of storage and use of information and human biological material in research and to learn the hybrid composition that followed norms and patterns for these techno-scientific practices.

The past two decades have witnessed rapid growth in the amount of data available to us. Many fields, including physics, biology, and medical studies, generate enormous datasets with a large sample size, a high number of dimensions, or both. For example, some datasets in physics contains millions of records. It is forecasted by Statista Survey that in 2022, there will be over 86 millions users of health apps in United States, which will generate massive mHealth data. In addition, more and more large studies have been carried out, such as the UK Biobank study. This gives us unprecedented access to data and allows us to extract and infer vital information. Meanwhile, it also poses new challenges for statistical methodologies and computational algorithms. For increasingly large datasets, computation can be a big hurdle for valid analysis. Conventional statistical methods lack the scalability to handle such large sample size. In addition, data storage and processing might be beyond usual computer capacity. The UK Biobank genotypes and phenotypes dataset contains about 500,000 individuals and more than 800,000 genotyped single nucleotide polymorphism (SNP) measurements per person, the size of which may well exceed a computer's physical memory. Further, the high dimensionality combined with the large sample size could lead to heavy computational cost and algorithmic instability. The aim of this dissertation is to provide some statistical approaches to address the issues. Chapter 1 provides a review on existing literature. In Chapter 2, a novel perturbation subsampling approach is developed based on independent and identically distributed stochastic weights for the analysis of large scale data. The method is justified based on optimizing convex criterion functions by establishing asymptotic consistency and normality for the resulting estimators. The method can provide consistent point estimator and variance estimator simultaneously. The method is also feasible for a distributed framework. The finite sample performance of the proposed method is examined through simulation studies and real data analysis. In Chapter 3, a repeated block perturbation subsampling is developed for the analysis of large scale longitudinal data using generalized estimating equation (GEE) approach. The GEE approach is a general method for the analysis of longitudinal data by fitting marginal models. The proposed method can provide consistent point estimator and variance estimator simultaneously. The asymptotic properties of the resulting subsample estimators are also studied. The finite sample performances of the proposed methods are evaluated through simulation studies and mHealth data analysis. With the development of technology, large scale high dimensional data is also increasingly prevailing. Conventional statistical methods for high dimensional data such as adaptive lasso (AL) lack the scalability to handle processing of such large sample size. Chapter 4 introduces the repeated perturbation subsampling adaptive lasso (RPAL), a new procedure which incorporates features of both perturbation and subsampling to yield a robust, computationally efficient estimator for variable selection, statistical inference and finite sample false discovery control in the analysis of big data. RPAL is well suited to modern parallel and distributed computing architectures and furthermore retains the generic applicability and statistical efficiency. The theoretical properties of RPAL are studied and simulation studies are carried out by comparing the proposed estimator to the full data estimator and traditional subsampling estimators. The proposed method is also illustrated with the analysis of omics datasets.

Ce mémoire de maîtrise tente de répondre à la question suivante : quelle est la valeur de l’autonomie dans le domaine de la recherche biomédicale, et plus particulièrement dans le contexte des biobanques? Pour éclairer cette question, nous étudierons le concept d’autonomie dans deux domaines d’éthique appliquée : l’éthique de la recherche et l’éthique de la santé publique. Depuis la deuxième moitié du XXe siècle, le respect de l’autonomie individuelle est un principe programmatique de l’éthique de la recherche. Or, sa conceptualisation a été accusée d’être trop individualiste pour répondre aux enjeux qui caractérisent la recherche en santé d’aujourd’hui. En éthique de la santé publique, au contraire, l’autonomie individuelle prend une place moins prépondérante et fait droit à des dimensions sociopolitiques de l’autonomie qui ont été négligées par le domaine de l’éthique de la recherche. Ces dimensions mettent en lumière le rôle des institutions dans la promotion de l’autonomie tout comme l’importance pour les individus d’avoir la possibilité de participer à l’élaboration des politiques de santé qui les concernent pour réaliser leur autonomie. Le cas de la recherche en contexte de biobanques sera l’occasion d’appliquer ces différentes conceptualisations de l’autonomie à un domaine de recherche en pleine expansion qui se situe au carrefour de l’éthique de la recherche et de l’éthique de la santé publique. / This master's thesis attempts to answer the following question: what is the value of autonomy in biomedical research, and more particularly in the context of biobanks? To answer this question, we will study the concept of autonomy in two areas of applied ethics: research ethics and public health ethics. Since the second half of the 20th century, respect for individual autonomy has been a programmatic principle of research ethics. However, many authors consider this conceptualization to be too individualistic to respond to the challenges that characterize health research today. In public health ethics, on the contrary, individual autonomy takes a less preponderant role and acknowledges some socio-political dimensions of autonomy that have been neglected by the field of research ethics. These dimensions highlight the role of institutions in promoting autonomy as well as the need for individuals to have the opportunity to participate in the development of health policies in order to achieve their autonomy. Finally, the case study of biobanks will provide an opportunity to apply these different conceptualizations of autonomy to a rapidly expanding field of research located at the crossroads between research ethics and public health.

L'objectif de mon travail de thèse était de développer des outils bio informatiques permettant d'améliorer la traditionnelle gestion de l'information scientifique au sein d'un grand centre de recherche et en particulier au sein d'une plateforme de génomique. Trois outils ont été développés: un cahier de laboratoire électronique, un système de gestion de l'information de laboratoire pour des applications de génomique dont le séquençage de nouvelle génération, ainsi qu'un système de gestion des échantillons pour de grandes bio-banques. Ce travail a été réalisé en étroite collaboration avec des biologistes, épidémiologistes et informaticiens. Il a également inclus la mise en place d'interactions entre les différents outils pour former un système informatique intégré. Les trois outils ont été rapidement adoptés par l'ensemble des scientifiques du centre de recherche et sont désormais utilisés au quotidien pour le suivi de toutes les activités de laboratoire mais aussi plus globalement pour les autres activités scientifiques du centre de recherche. Ces outils sont transposables dans d'autres instituts de recherche / The aim of my thesis work was to develop bioinformatics tools to improve the traditional scientific information management within a large research centre and especially within a genomics platform. Three tools have been developed: an electronic laboratory notebook, a laboratory information management system for genomics applications including next generation sequencing, as well as a sample management system for large biobanks. This work has been conducted in close collaboration with biologists, epidemiologists and IT specialists. It has also included the setup of interactions between the different tools to make an integrated IT system. The three tools have been rapidly adopted by all the scientists of the research centre and are now daily used for the tracking of all the laboratory’s activities but also more globally for the research centre’s other scientific activities. These tools are transposable in other research institutes

La pharmacogénomique (PGx) étudie le concept selon lequel les déterminants génétiques peuvent aider à prédire la réponse clinique d’un patient aux médicaments. Les concentrations plasmatiques de ces derniers sont essentielles pour déterminer l’exposition, les profils pharmacocinétiques (PK), les effets cliniques et éventuellement les doses des médicaments, dont la plupart sont métabolisés par des enzymes hépatiques, les cytochromes P450 (CYPs). Néanmoins, la plupart des découvertes en matière de PGx concernant la prédiction des profils de concentrations des médicaments ont généralement recours à des plans d’études PK traditionnels avec une approche fonctionnelle. Bien qu’utile, cette méthodologie comporte des limites pour les études PGx, notamment le nombre restreint de sujets inclus, qui réduit la puissance statistique des associations PGx et limite l’identification de nouveaux variants génétiques moins fréquents. À l’inverse, les grandes cohortes observationnelles sont largement utilisées pour identifier des marqueurs génétiques physiopathologiques. Cette thèse de doctorat visait donc à 1) synthétiser les données publiées concernant les effets cliniques des polymorphismes génétiques de l’enzyme CYP2D6 sur le traitement au métoprolol, un agent β-bloquant. Les concentrations plasmatiques de métoprolol ont montré à plusieurs reprises qu’elles étaient fortement influencées par la PGx du CYP2D6; 2) développer une nouvelle méthode bioanalytique capable de quantifier les concentrations chirales de métoprolol des patients dans un contexte clinique; 3) mener une étude clinique en utilisant une grande cohorte observationnelle, ou biobanque, comme preuve de concept pour recréer l’association précédemment établie entre les phénotypes inférés des génotypes du CYP2D6 et les concentrations plasmatiques de métoprolol. Ces projets sont présentés en tant que chapitres de thèse et sous forme de manuscrits publiés. Le premier projet consistait en une revue systématique qui a permis d’extraire toutes les études relatives à la PGx du métoprolol-CYP2D6. La synthèse qualitative a suggéré que les métaboliseurs lents du CYP2D6, dépourvus de capacité enzymatique, avaient des valeurs plus élevées concernant les réductions de la fréquence cardiaque et de tension artérielle, ainsi que la survenue d’épisodes bradycardiques relativement aux autres phénotypes. Une méta-analyse ultérieure a confirmé la significativité de ces associations. Le deuxième projet a combiné des techniques bioanalytiques telles que la dérivation, l’extraction en phase solide et la chromatographie liquide avec spectrométrie de masse en tandem. Une méthode permettant de surmonter les limites analytiques antérieures a été validée avec succès pour mesurer les concentrations plasmatiques de (S)-métoprolol, l’énantiomère pharmacologiquement actif, et de son métabolite spécifique au CYP2D6. L’applicabilité d’une telle méthode a ensuite été démontrée grâce aux échantillons d’un groupe de patients issus de la Cohorte Hospitalière de l’Institut de Cardiologie de Montréal (ICM). Puis, le troisième projet présente la réalisation de l’étude LEVEL-PGx (LEVEraging Large observational cohort studies to identify pharmacogenetic determinants of drug dosing : A proof-of-concept study in the Montreal Heart Institute Hospital Cohort). L’étude portait sur un échantillon de >1000 patients sélectionnés dans la cohorte hospitalière de l’ICM, incluant leur génotypage pour CYP2D6 et la quantification du métoprolol racémique et de son métabolite spécifique au CYP2D6 dans des échantillons provenant de la Biobanque de l’ICM. Un seul échantillon unique et aléatoire par patient a été utilisé. Le recours à des modèles multivariables a validé le concept selon lequel de grandes cohortes transversales recueillant des échantillons biologiques pouvaient être utilisées afin d’identifier des associations PGx de concentrations de médicaments et ce, à des valeurs satisfaisant les seuils de significativité d’essais pangénomiques. D’autres analyses de cette cohorte ont indiqué que cette méthodologie parvenait à identifier des associations PGx qui influençaient la fréquence cardiaque au repos et la posologie du métoprolol à-travers les phénotypes du CYP2D6 et pour les déterminants génétiques uniques, même en présence de co-médications. Cependant, ces associations PGx avec les paramètres cliniques n’ont pas atteint une significativité applicable aux seuils pangénomiques. En résumé, par la reproduction d’une association PGx préalablement démontrée, l’ensemble des travaux présentés dans cette thèse suggère que l’identification et la découverte de nouveaux déterminants génétiques prédictifs des concentrations et des doses des médicaments pourrait s’effectuer par le biais de grandes cohortes observationnelles à l’échelle du génome. Ces approches permettraient de développer des modèles prédictifs plus précis de l’exposition et de la réponse aux médicaments, ce qui pourrait favoriser les découvertes PGx et, dans certains cas, éventuellement développer le potentiel translationnel d’une approche thérapeutique personnalisée selon le profil génétique des patients. / Pharmacogenomics (PGx) studies the concept that genetic determinants can help predict a patient’s clinical response to therapies. Drug concentrations are an essential component to determining the exposure, pharmacokinetic (PK) profiles, clinical effects, and potentially drug doses, most of which are metabolized through the cytochrome P450 (CYPs) liver enzymes. Nevertheless, most PGx discoveries regarding the prediction of drug concentration profiles have generally resorted to traditional PK study designs with a functional approach. Though useful, this methodology contains limitations for gene-drug interaction studies, most notably the restricted number of subjects included, which reduces the statistical power for PGx associations and limits the identification of new, less frequent genetic variants. On the opposite, large observational cohorts have long been utilized for identifying genetic markers of disease. This doctoral thesis therefore aimed to 1) synthesize published data regarding the clinical effects of CYP2D6 genetic polymorphism on metoprolol therapy. A β-blocker, metoprolol plasma concentrations have shown repeatedly to be heavily influenced by the PGx of the CYP2D6 enzyme; 2) develop a new bioanalytical method able to quantify patients’ chiral concentrations of metoprolol in a clinical setting; 3) conduct a clinical study using a large observational cohort, or biobank, as a proof of concept to recreate the previously established association between CYP2D6 genotype-inferred phenotypes and metoprolol plasma concentrations. Those projects are presented as thesis chapters in the form of published manuscripts. The first project was a systematic review that allowed us to find all studies pertaining to the PGx of metoprolol. The qualitative synthesis suggested that CYP2D6 poor metabolizers (PMs), without enzymatic capacity, had greater values regarding reductions in heart rate, blood pressures, and occurrences in bradycardia relative to non-PMs. A subsequent meta-analysis confirmed the significance of those associations. The second project combined bioanalytical techniques such as derivatization, solid phase extraction, and liquid chromatography-tandem mass spectrometry. A method overcoming previous analytical shortcomings was successfully validated to measure (S)-metoprolol plasma concentrations and its CYP2D6-specific metabolite. Its application was later demonstrated in a group of patients from the Montreal Heart Institute (MHI) Hospital Cohort. Then, the third project presents the conduct of the LEVEL-PGx study (LEVEraging Large observational cohort studies to identify pharmacogenetic determinants of drug dosing: A proof-of-concept study in the Montreal Heart Institute Hospital Cohort). The study implicated a sample of >1000 selected patients selected from the MHI Hospital Cohort, along with the genotyping of CYP2D6, and the quantification of racemic metoprolol and its CYP2D6-specific metabolite in samples from the MHI Biobank. A single, random sample per patient was used. Multivariable modeling validated the concept that large observational cohorts collecting biospecimens could be utilized to identify PGx associations of drug concentrations with genome-wide significance. Further analyses in our cohort indicated that the tested PGx associations influenced resting heart rate and metoprolol daily drug dosage across CYP2D6 phenotypes and for single genetic determinants, regardless of interfering comedications. However, such PGx associations with clinical parameters could not achieve genome-wide significance. In summary, the body of work presented in this thesis suggested that, using a previously validated PGx association, the identification of novel genetic determinants predictive of drug concentrations and dosage could be discovered and identified at the genome-wide level with large observational cohorts. These approaches would help develop more accurate predictive models of drug exposure and response, which could favor PGx discoveries and the translational potential of a personalized approach to treatments according to a patient’s genetic profile.