No description available.

No description available.

Doctor of Philosophy / Department of Electrical and Computer Engineering / Sanjoy Das / Stochastic algorithms are widely used in various modeling and optimization problems. Evolutionary algorithms are one class of population-based stochastic approaches that are inspired from Darwinian evolutionary theory. A population of candidate solutions is initialized at the first generation of the algorithm. Two variation operators, crossover and mutation, that mimic the real world evolutionary process, are applied on the population to produce new solutions from old ones. Selection based on the concept of survival of the fittest is used to preserve parent solutions for next generation. Examples of such algorithms include genetic algorithm (GA) and genetic programming (GP). Nevertheless, other stochastic algorithms may be inspired from animals’ behavior such as particle swarm optimization (PSO), which imitates the cooperation of a flock of birds. In addition, stochastic algorithms are able to address multi-objective optimization problems by using the concept of dominance. Accordingly, a set of solutions that do not dominate each other will be obtained, instead of just one best solution. This thesis proposes a multi-objective GP-PSO hybrid algorithm to recover gene regulatory network models that take environmental data as stimulus input. The algorithm infers a model based on both phenotypic and gene expression data. The proposed approach is able to simultaneously infer network structures and estimate their associated parameters, instead of doing one or the other iteratively as other algorithms need to. In addition, a non-dominated sorting approach and an adaptive histogram method based on the hypergrid strategy are adopted to address ‘convergence’ and ‘diversity’ issues in multi-objective optimization. Gene network models obtained from the proposed algorithm are compared to a synthetic network, which mimics key features of Arabidopsis flowering control system, visually and numerically. Data predicted by the model are compared to synthetic data, to verify that they are able to closely approximate the available phenotypic and gene expression data. At the end of this thesis, a novel breeding strategy, termed network assisted selection, is proposed as an extension of our hybrid approach and application of obtained models for plant breeding. Breeding simulations based on network assisted selection are compared to one common breeding strategy, marker assisted selection. The results show that NAS is better both in terms of breeding speed and final phenotypic level.

L’organogenèse post-embryonnaire des racines latérales joue un rôle essentiel dans l’établissement de l’architecture du système racinaire des plantes, et donc dans leur croissance et leur performance. L’objectif de cette thèse est de caractériser le réseau de gènes régulant le développement des racines latérales et en particulier, l’organisation fonctionnelle du primordium de racine latérale, formant un nouveau méristème racinaire, chez la plante modèleArabidopsis thaliana en combinant des études de biologie des systèmes appliquées à la dynamique du transcriptome lors de la formation des racines latérales avec la caractérisation fonctionnelle de gènes candidats pour la régulation de ce phénomène d’organogenèse.La première partie de la thèse concerne l’identification des cibles de PUCHI, un facteur de transcription de type AP2/EREBP impliqué dans le contrôle de la prolifération et de la différentiation cellulaire dans le primordium de racine latérale. Le phénotype liés à la parte de fonction de PUCHI a été caractérisé en détail et à mis en évidence un rôle de ce facteur de transcription dans l'initiation des racines latérales et le développement et l'organisation des primordia. Par l’analyse de profils spatiaux et temporels d’expression de gènes, nous avons pu mettre en évidence que l’expression de gènes codant des protéines impliquées dans la biosynthèse des acides gras à très longues chaînes (VLCFA) est transitoirement activée durant la formation de la racine latérale et que cette dynamique est dépendante de PUCHI. De plus, le mutant kcs1-5, perturbé dans la biosynthèse de VLCFAs, présente un phénotype de développement des racines latérales similaire à celui de puchi-1. Par ailleurs, la perte de fonction puchi-1 augmente fortement la formation de cals continus dans des racines cultivées sur milieu inducteur riche en auxine, ce qui est cohérent avec le rôle récemment décrit des VLCFA racinaires dans la formation et l’organisation de cals distincts lorsque la racine est cultivé sur milieu inducteur de cals. L'ensemble de nos résultats démontre que PUCHI régule positivement l’expression de gènes de biosynthèse de VLCFAs lors de la formation de racines latérales et la callogenèse. Nos résultats confortent également l’hypothèse selon laquelle la formation des racines latérales et celle de cals racinaires partagent des mécanismes de régulation communs.La seconde partie de la thèse s’intéresse à l’identification de facteurs régulateurs clés dans l’organisation fonctionnelle du primordium de racine latérale et particulièrement, l’organisation d’un nouveau méristème racinaire. J’ai contribué à produire de nouvelles lignées de plantes permettant de suivre en temps réel par microscopie confocale la mise en place des identités cellulaires caractéristiques d’un méristème racinaire dans le primordium de racine latérale en développement. En utilisant un algorithme d’inférence de réseau de gènes, j’ai produit puis analysé les relations prédites de régulation entre gènes d’intérêt, afin d’identifier des gènes candidats potentiellement impliqués dans la formation du centre quiescent, un élément clé dans l’organisation du primordium et la mise en place du nouveau méristème racinaire. La caractérisation fonctionnelle de certains de ces gènes candidats a été initiée.Ces travaux de thèse ont donc contribué à mieux comprendre les mécanismes de régulation de la formation des racines latérales chez Arabidopsis thaliana. / Post-embryonic lateral root organogenesis plays an essential role in defining plant root system architecture, and therefore plant growth and fitness. The aim of the thesis is to elucidate the gene regulatory network regulating lateral root development and de novo root meristem formation during root branching in the model plant Arabidopsis thaliana by combining a system-biology based analysis of lateral root primordium transcritome dynamics with the functional characterization of genes possibly involved in regulating lateral root organogenesis.The first part of the thesis deals with the identification the target genes of PUCHI, an AP2/EREBP transcription factor that is involved in controlling cell proliferation and differentiation during lateral root formation. We showed that loss of PUCHI function leads to defects lateral root initiation and primordium growth and organisation. We found that several genes coding for proteins of the very long chain fatty acid (VLCFA) biosynthesis machinery are transiently induced in a PUCHI-dependent manner during lateral root development. Moreover, a mutant perturbed in VLCFA biosynthesis (kcs1-5) displays similar lateral root development defects as does puchi-1. In addition, puchi-1 loss of function mutant roots show enhanced and continuous callus formation in auxin-rich callus induction medium, consistent with the recently reported role of VLCFAs in organizing separated callus proliferation on this inductive growing medium. Thus, our results show that PUCHI positively regulates the expression of VLCFA biosynthesis genes during lateral root development, and further support the hypothesis that lateral root and callus formation share common genetic regulatory mechanisms.A second part of the thesis specifically addresses the issue of identifying key regulators of root meristem organization in the developing lateral root primordium. Material enabling the tracking of meristem cell identity establishment in developing primordia with live confocal microscopy was generated. A gene network inference was run to predict potential regulatory relationships between genes of interest during the time course of lateral root development. It identified potential regulators of quiescent center formation, a key step in functional organization of the lateral root primordia into a new root apical meristem. The characterization of some of these candidate genes was initiated.Altogether, this work participated in deciphering the genetic regulation of lateral root formation in Arabidopsis thaliana .

Genes encode proteins, some of which in turn regulate other genes. Such interactions make up gene regulatory relationships or (dynamic) gene regulatory networks. With advances in the measurement technology for gene expression and in genome sequencing, it has become possible to measure the expression level of thousands of genes simultaneously in a cell at a series of time points over a specific biological process. Such time-course gene expression data may provide a snapshot of most (if not all) of the interesting genes and may lead to a better understanding gene regulatory relationships and networks. However, inferring either gene regulatory relationships or networks puts a high demand on powerful computational methods that are capable of sufficiently mining the large quantities of time-course gene expression data, while reducing the complexity of the data to make them comprehensible. This dissertation presents several computational methods for inferring gene regulatory relationships and gene regulatory networks from time-course gene expression. These methods are the result of the authors doctoral study. Cluster analysis plays an important role for inferring gene regulatory relationships, for example, uncovering new regulons (sets of co-regulated genes) and their putative cis-regulatory elements. Two dynamic model-based clustering methods, namely the Markov chain model (MCM)-based clustering and the autoregressive model (ARM)-based clustering, are developed for time-course gene expression data. However, gene regulatory relationships based on cluster analysis are static and thus do not describe the dynamic evolution of gene expression over an observation period. The gene regulatory network is believed to be a time-varying system. Consequently, a state-space model for dynamic gene regulatory networks from time-course gene expression data is developed. To account for the complex time-delayed relationships in gene regulatory networks, the state space model is extended to be the one with time delays. Finally, a method based on genetic algorithms is developed to infer the time-delayed relationships in gene regulatory networks. Validations of all these developed methods are based on the experimental data available from well-cited public databases.

Genes encode proteins, some of which in turn regulate other genes. Such interactions make up gene regulatory relationships or (dynamic) gene regulatory networks. With advances in the measurement technology for gene expression and in genome sequencing, it has become possible to measure the expression level of thousands of genes simultaneously in a cell at a series of time points over a specific biological process. Such time-course gene expression data may provide a snapshot of most (if not all) of the interesting genes and may lead to a better understanding gene regulatory relationships and networks. However, inferring either gene regulatory relationships or networks puts a high demand on powerful computational methods that are capable of sufficiently mining the large quantities of time-course gene expression data, while reducing the complexity of the data to make them comprehensible. This dissertation presents several computational methods for inferring gene regulatory relationships and gene regulatory networks from time-course gene expression. These methods are the result of the authors doctoral study. Cluster analysis plays an important role for inferring gene regulatory relationships, for example, uncovering new regulons (sets of co-regulated genes) and their putative cis-regulatory elements. Two dynamic model-based clustering methods, namely the Markov chain model (MCM)-based clustering and the autoregressive model (ARM)-based clustering, are developed for time-course gene expression data. However, gene regulatory relationships based on cluster analysis are static and thus do not describe the dynamic evolution of gene expression over an observation period. The gene regulatory network is believed to be a time-varying system. Consequently, a state-space model for dynamic gene regulatory networks from time-course gene expression data is developed. To account for the complex time-delayed relationships in gene regulatory networks, the state space model is extended to be the one with time delays. Finally, a method based on genetic algorithms is developed to infer the time-delayed relationships in gene regulatory networks. Validations of all these developed methods are based on the experimental data available from well-cited public databases.

To understand how the components of a complex system like the biological cell interact and regulate each other, we need to collect data for how the components respond to system perturbations. Such data can then be used to solve the inverse problem of inferring a network that describes how the pieces influence each other. The work in this thesis deals with modelling the cell regulatory system, often represented as a network, with tools and concepts derived from systems biology. The first investigation focuses on network sparsity and algorithmic biases introduced by penalised network inference procedures. Many contemporary network inference methods rely on a sparsity parameter such as the L1 penalty term used in the LASSO. However, a poor choice of the sparsity parameter can give highly incorrect network estimates. In order to avoid such poor choices, we devised a method to optimise the sparsity parameter, which maximises the accuracy of the inferred network. We showed that it is effective on in silico data sets with a reasonable level of informativeness and demonstrated that accurate prediction of network sparsity is key to elucidate the correct network parameters. The second investigation focuses on how knowledge from association networks can be transferred to regulatory network inference procedures. It is common that the quality of expression data is inadequate for reliable gene regulatory network inference. Therefore, we constructed an algorithm to incorporate prior knowledge and demonstrated that it increases the accuracy of network inference when the quality of the data is low. The third investigation aimed to understand the influence of system and data properties on network inference accuracy. L1 regularisation methods commonly produce poor network estimates when the data used for inference is ill-conditioned, even when the signal to noise ratio is so high that all links in the network can be proven to exist for the given significance. In this study we elucidated some general principles for under what conditions we expect strongly degraded accuracy. Moreover, it allowed us to estimate expected accuracy from conditions of simulated data, which was used to predict the performance of inference algorithms on biological data. Finally, we built a software package GeneSPIDER for solving problems encountered during previous investigations. The software package supports highly controllable network and data generation as well as data analysis and exploration in the context of network inference. / <p>At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 4: Manuscript.</p><p> </p>

In biology, regulatory networks are sets of macromolecules, mostly proteins and RNAs that interact to execute task. The main players in regulatory networks are DNAbinding proteins, also called transcription factors as they modulate the first step in gene expression. A gene regulatory network (GRN) is a set of genes or proteins that interact with each other to control a specific cell function. Gene regulatory networks are important in development, differentiation and to respond to environmental cues. Gene regulatory networks (GRNs) are the on-off switches of a cell operating at the gene and/or protein level. The modeling methods can be broadly categorized into continuous and discrete. In this work , we dedicate attention to discrete models on cell senescence models for Astrocyte [35], the modelling of drug synergies to control gastric cancer [38], and we also wrote a paper about Discrete and Continuous Model, advantage or disadvantage of these models and a list of available softwares for using these kind of approaches. / Em biologia, redes regulatórias são conjuntos de macromoléculas, principalmente proteínas e RNAs que interagem para executar uma tarefa. As proteínas de ligação de DNA, também chamadas de fatores de transcrição, são as principais executoras nas redes regulatórias, visto que modulam o primeiro passo na expressão gênica. Uma rede genética regulatória (RRG) é um conjunto de genes ou proteínas que interagem uns com os outros para controlar uma função celular específica. Redes regulatórias são importantes no desenvolvimento, diferenciação e para responder aos sinais ambientais. Elas são os botões de liga/desliga de uma célula operando no nível do gene e/ou proteína. Seus métodos de modelagem podem ser geralmente classificados em contínuos e discretos. Neste trabalho, dedicamos atenção aos modelos discretos em senescência celular para astrócitos [35], a modelagem de sinergias de drogas para controle do câncer gástrico [38] e também escrevemos um artigo sobre Modelos Discretos e Contínuos, vantagens e desvantagens desses modelos e listagem dos softwares disponíveis para uso nesse tipo de abordagem.

Comprendre comment l’expression des gènes est régulée spécifiquement dans chaque tissu et de manière dynamique au cours du temps demeure une étape centrale de notre compréhension de l’organogénèse. L’identification des éléments cis-régulateurs de la transcription de manière tissu-spécifique peut permettre de comprendre les règles logiques d’organisation du réseau de gènes régulateur et aussi d’identifier de nouveaux acteurs (facteurs de transcription notamment). L’analyse de marques de chromatine (H3K27ac et H3K4me3) spécifiquement dans les cardioblastes (104 cellules) au cours de la différentiation a permis l’identification en masse de régions cis-régulatrices de la transcription. Via une approche d’apprentissage, de nouvelles régions régulatrices spécifiques des cardiomyocytes ainsi que 2 nouveaux facteurs de transcription (bagpipe, hamlet) ont été identifiées. L’alignement multiple des régions régulatrices suggère que les régions associées à H3K27ac dans les cellules cardiaques durant ces étapes de l’organogénèse partagent une séquence consensus. Ces nouveaux éléments régulateurs viennent compléter le réseau de gène régulateur au cours des étapes tardives de la cardiogénèse. / Understanding how gene expression is spatio-temporally regulated remains a crucial step in our understanding of organogenesis. Identification of transciptional cis-regulatory elements in a tissu-specific manner could allow to understand logical rules leading regulatory network organisation and to identify new actors (in particular transcription factors). Analysis of chromatin marks (H3K27ac and H3K4me3) specifically in cardiac cells (104 cells) during differentiation allowed the identification of transcriptional cis-regulatory regions. Via a machine learning approach, new cardiac specific regulatory regions and two transcription factors (bagpipe and hamlet) have been identified. Multiple sequence alignment of regulatory regions suggests that regions associated to H3K27ac in cardiac cells during these steps of organogenesis share a consensus sequence. These new regulatory elements integrate and complete the gene regulatory network underlying late steps of cardiogenesis.

Gene expression regulation is dynamic and specific to various factors such as developmental stages, environmental conditions, and stimulation of pathogens. Nowadays, a tremendous amount of transcriptome data sets are available from diverse species. This trend enables us to perform comparative transcriptome analysis that identifies conserved or diverged gene expression responses across species using transcriptome data. The goal of this dissertation is to develop and apply approaches of comparative transcriptomics to transfer knowledge from model species to non-model species with the hope that such an approach can contribute to the improvement of crop yield and human health. First, we presented a comprehensive method to identify cross-species modules between two plant species. We adapted the unsupervised network-based module finding method to identify conserved patterns of co-expression and functional conservation between Arabidopsis, a model species, and soybean, a crop species. Second, we compared drought-responsive genes across Arabidopsis, soybean, rice, corn, and Populus in order to explore the genomic characteristics that are conserved under drought stress across species. We identified hundreds of common gene families and conserved regulatory motifs between monocots and dicots. We also presented a BLS-based clustering method which takes into account evolutionary relationships among species to identify conserved co-expression genes. Last, we analyzed single-cell RNA-seq data from monocytes to attempt to understand regulatory mechanism of innate immune system under low-grade inflammation. We identified novel subpopulations of cells treated with lipopolysaccharide (LPS), that show distinct expression patterns from pro-inflammatory genes. The data revealed that a promising therapeutic reagent, sodium 4-phenylbutyrate, masked the effect of LPS. We inferred the existence of specific cellular transitions under different treatments and prioritized important motifs that modulate the transitions using feature selection by a random forest method. There has been a transition in genomics research from bulk RNA-seq to single-cell RNA-seq, and scRNA-seq has become a widely used approach for transcriptome analysis. With the experience we gained by analyzing scRNA-seq data, we plan to conduct comparative single-cell transcriptome analysis across multiple species. / Doctor of Philosophy / All cells in an organism have the same set of genes, but there are different cell types, tissues, organs with different functions as the organism ages or under different conditions. Gene expression regulation is one mechanism that modulates complex, dynamic, and specific changes in tissues or cell types for any living organisms. Understanding gene regulation is of fundamental importance in biology. With the rapid advancement of sequencing technologies, there is a tremendous amount of gene expression data (transcriptome) from individual species in public repositories. However, major studies have been reported from several model species and research on non-model species have relied on comparison results with a few model species. Comparative transcriptome analysis across species will help us to transform knowledge from model species to non-model species and such knowledge transfer can contribute to the improvement of crop yields and human health. The focus of my dissertation is to develop and apply approaches for comparative transcriptome analysis that can help us better understand what makes each species unique or special, and what kinds of common functions across species have been passed down from ancestors (evolutionarily conserved functions). Three research chapters are presented in this dissertation. First, we developed a method to identify groups of genes that are commonly co-expressed in two species. We chose seed development data from soybean with the hope to contribute to crop improvement. Second, we compared gene expression data across five plant species including soybean, rice, and corn to provide new perspectives about crop plants. We chose drought stress to identify conserved functions and regulatory factors across species since drought stress is one of the major stresses that negatively impact agricultural production. We also proposed a method that groups genes with evolutionary relationships from an unlimited number of species. Third, we analyzed single-cell RNA-seq data from mouse monocytes to understand the regulatory mechanism of the innate immune system under low-grade inflammation. We observed how innate immune cells respond to inflammation that could cause no symptoms but persist for a long period of time. Also, we reported an effect of a promising therapeutic reagent (sodium 4-phenylbutyrate) on chronic inflammatory diseases. The third project will be extended to comparative single-cell transcriptome analysis with multiple species.