The recent technology of single-cell RNA sequencing can be used to investigate molecular, transcriptional, changes in cells as they develop. I reviewed the literature on the technology, and made a large scale quantitative comparison of the different implementations of single cell RNA sequencing to identify their technical limitations. I investigate how to model transcriptional changes during cellular development. The general forms of expression changes with respect to development leads to nonparametric regression models, in the forms of Gaussian Processes. I used Gaussian process models to investigate expression patterns in early embryonic development, and compared the development of mice and humans. When using in vivo systems, ground truth time for each cell cannot be known. Only a snapshot of cells, all being in different stages of development can be obtained. In an experiment measuring the transcriptome of zebrafish blood precursor cells undergoing the development from hematopoietic stem cells to thrombocytes, I used a Gaussian Process Latent Variable model to align the cells according to the developmental trajectory. This way I could investigate which genes were driving the development, and characterise the different patterns of expression. With the latent variable strategy in mind, I designed an experiment to study a rare event of murine embryonic stem cells entering a state similar to very early embryos. The GPLVM can take advantage of the nonlinear expression patterns involved with this process. The results showed multiple activation events of genes as cells progress towards the rare state. An essential feature of cellular biology is that precursor cells can give rise to multiple types of progenitor cells through differentiation. In the immune system, naive T-helper cells differentiate to different sub-types depending on the infection. For an experiment where mice were infected by malaria, the T-helper cells develop into two cell types, Th1 and Tfh. I model this branching development using an Overlapping Mixture of Gaussian Processes, which let me identify both which cells belong to which branch, and learn which genes are involved with the different branches. Researchers have now started performing high-throughput experiments where spatial context of gene expression is recorded. Similar to how I identify temporal expression patterns, spatial expression patterns can be identified nonparametrically. To enable researchers to make use of this technique, I developed a very fast method to perform a statistical test for spatial dependence, and illustrate the result on multiple data sets.

An estimated 170.38 million elderly adults suffer from some stage of age-related macular degeneration (AMD) worldwide, a vision defect that damages the macula, the central region of the retina required for sharp vision, such as reading, driving, and recognizing faces. Genetic factors strongly modify one's risk for developing AMD, and most of these genetic changes are found in genes of the alternative complement cascade, a component of the immune system. The lack of effective AMD prevention calls for the identification of druggable molecules and pathways. In my research, I use microarrays and RNA sequencing to investigate the events occurring in early AMD, the reasons for macular susceptibility to AMD, and the events triggering aberrant blood vessel growth in late AMD. First, I found that genes associated with endothelial cells tend to be expressed at lower levels in human donors eyes affected by early AMD than in control eyes, concordant with previous studies indicating loss of choriocapillaris in early AMD. Second, I found that molecular signals across regions of the retina, retinal pigment epithelium, and choroid generally mirror the distribution of cell types in these regions. Third, I found that damage to cultured primate chorioretinal endothelial cells by the end product of complement activation, membrane attack complex, produces an environment conducive to choroidal neovascularization, a symptom of late-stage AMD. I propose a model that bridges genetic variants in the complement cascade genes with blood vessel loss in early AMD and the pathological growth of blood vessels in late AMD.

Shortcomings of current methods of prostate cancer detection draw attention to a need for improved biomarkers. The TMPRSS2:ERG gene fusion leads to the overexpression of ERG, an ETS family transcription factor, and is the most prevalent genetic lesion in prostate cancer, but its clinical utility remains to be defined. Two radical prostatectomy samples were analysed by next-generation whole transcriptome sequencing. The chosen samples differed in fusion gene status, as previously determined by RT-PCR. The involvement of novel and previously reported prostate cancer-related transcripts, Wnt signalling, p53 effector loss and several ETS-regulated pathways was identified in the prostate cancer cases examined. ERG was found to directly transactivate RhoGDIB, a gene associated with fusion-positive prostate cancer. Overexpression of RhoGDIB elicited spindle-shaped morphology, faster cell migration and increased cell proliferation, phenotypic changes suggestive of cancer progression. The present findings confirm the value of comprehensive sequencing for biomarker development and indicate avenues of future study.

Shortcomings of current methods of prostate cancer detection draw attention to a need for improved biomarkers. The TMPRSS2:ERG gene fusion leads to the overexpression of ERG, an ETS family transcription factor, and is the most prevalent genetic lesion in prostate cancer, but its clinical utility remains to be defined. Two radical prostatectomy samples were analysed by next-generation whole transcriptome sequencing. The chosen samples differed in fusion gene status, as previously determined by RT-PCR. The involvement of novel and previously reported prostate cancer-related transcripts, Wnt signalling, p53 effector loss and several ETS-regulated pathways was identified in the prostate cancer cases examined. ERG was found to directly transactivate RhoGDIB, a gene associated with fusion-positive prostate cancer. Overexpression of RhoGDIB elicited spindle-shaped morphology, faster cell migration and increased cell proliferation, phenotypic changes suggestive of cancer progression. The present findings confirm the value of comprehensive sequencing for biomarker development and indicate avenues of future study.

Bacterial pathogens rely on precise regulation of gene expression to coordinate host infection processes and resist invasion by mobile genetic elements. An interconnected network of protein and RNA regulators dynamically controls the expression of virulence factors using a variety of mechanisms. In this thesis, the role of selected regulators, belonging to the class of small RNAs (sRNAs), is investigated. Streptococcus pyogenes is a pathogen responsible for a wide range of human diseases. Genome-wide screenings have indicated that S. pyogenes encodes numerous sRNAs, yet only a limited number have been characterized. A major goal of this study was to identify and characterize novel sRNAs and antisense RNAs (asRNAs) using RNA sequencing analysis. We validated 30 novel sRNAs and asRNAs, and identified 9 sRNAs directly cleaved by the ribonucleases RNase III and/or RNase Y. Previous work from the laboratory has highlighted the role of sRNAs from the type II Clustered Regularly Interspaced Short Palindromic Repeats-CRISPR associated proteins (CRISPR-Cas) systems in S. pyogenes. CRISPR-Cas systems provide adaptive immunity to prokaryotes against infection by mobile genetic elements. Two sRNAs, forming a complementary duplex (dual-RNA), are effectors of this system: the mature CRISPR RNAs (crRNAs) and the trans-activating crRNA (tracrRNA). The dual-RNA guides the Cas9 endonuclease to cleave both strands of the invading DNA in a sequence-specific manner. This RNA-programmable CRISPR-Cas9 system is now utilized for genome editing and engineering in a wide range of cells and organisms. To expand the potentialities of this tool, we both, searched for Cas9 orthologs and predicted numerous tracrRNA orthologs. We defined tracrRNA as a new family of sRNAs sharing the ability to base-pair to cognate crRNAs, without conservation of structure, sequence or location. We show that Cas9 and the dual tracrRNA:crRNAs are only interchangeable between closely related type II CRISPR-Cas systems. In summary, this thesis presents new insights into RNA-mediated regulatory mechanisms in S. pyogenes. We identified and described the expression of novel sRNAs, highlighting potential antisense RNAs. Focusing on the dual-RNA programmable type II CRISPR-Cas system, we provided evidence for co-evolution of the Cas9 enzyme with tracrRNA:crRNA, a basis for Cas9 multiplexing in genome editing.

The human body is made of billions of cells and nearly all have the same genome. However, there is a high diversity of cells, resulted from what part of the genome the cells use, i.e. which RNA molecules are expressed. Rapid advances within the field of sequencing allow us to determine the RNA molecules expressed in a specific cell at a certain time. The use of the new technologies has expanded our view of the human transcriptome and increased our understanding of when, where, and how each RNA molecule is expressed. The work presented in this thesis focuses on analysis of the human transcriptome. In Paper I, we describe an automated approach for sample preparation. This protocol was compared with the standard manual protocol, and we demonstrated that the automated version outperformed the manual process in terms of sample throughput while maintaining high reproducibility. Paper II addresses the impact of nuclear transcripts on gene expression. We compared total RNA from whole cells and from cytoplasm, showing that transcripts with long, structured 3’- and 5’-untranslated regions and transcripts with long protein coding sequences tended to be retained in the nucleus. This resulted in increased complexity of the total RNA fraction and fewer reads per unique transcript. Papers III and IV describe dynamics of the human muscle transcriptome. For Paper III, we systematically investigated the transcriptome and found remarkably high tissue homogeneity, however a large number of genes and isoforms were differentially expressed between genders. Paper IV describes transcriptome differences in response to repeated training. No transcriptome-based memory was observed, however a large number of isoforms and genes were affected by training. Paper V describes a transcript profiling protocol based on the method Reverse Transcriptase Multiplex Ligation-dependent Probe Amplification. We designed the method for a few selected transcripts whose expression patterns are important for detecting breast cancer cells, and optimized the method for single cell analysis. We successfully detected cells in human blood samples and applied the method to single cells, confirming the heterogeneity of a cell population. / Människokroppen är uppbyggd av miljarder celler och nästan alla innehåller samma arvsmassa. Trots detta finns det många olika celler med olika funktioner vilket är en följd av vilken del av arvsmassan som cellerna använder, dvs vilka RNA-molekyler som finns i varje cell. Den snabba utvecklingen av sekvenseringstekniker har gjort det möjligt att studera när, var och hur varje RNA-molekyl är uttryckt och att få en djupare förståelse för hur människans celler fungerar. Arbetet som presenteras i denna avhandling fokuserar på analys av RNA-molekyler i människans celler. I artikel I beskriver vi en automatiserad metod för att förbereda cellprov för RNA-sekvensering. Det automatiserade protokollet jämfördes med det manuella protokollet, och vi visade att det automatiserade protokollet överträffade det manuella när det gällde provkapacitet samtidigt som en höga reproducerbarheten behölls. I artikel II undersökte vi effekterna som RNA-molekyler från en del av cellen (cellkärnan) har på den totala mängden uttryckta RNA-molekyler. Vi jämförde RNA från hela cellen och från en del av cellen (cytoplasman) och visade att RNA-molekyler med långa och strukturerade 3'- och 5'-otranslaterade regioner och RNA-molekyler med långa proteinkodande sekvenser tenderade att hållas kvar i cellkärnan till en högre grad. Detta resulterade i en ökad komplexitet av RNA-molekylerna i hela cellen, medan vi i cytoplasma-fraktionen lättare kunde hitta de korta och svagt uttryckta RNA-molekyler. I Artikel III och IV studerar vi RNA-molekyler i människans skelettmuskler. I artikel III visar vi att andelen RNA-molekyler uttryckta i skelettmuskler är väldigt lika mellan muskler och mellan olika personer, men att ett stort antal RNA-molekyler var uttryckta i olika nivåer hos kvinnor och män. Artikel IV beskriver RNA-nivåer som svar på upprepade perioder av uthållighetsträning. Artikel V beskriver en metod för att studera ett fåtal utvalda RNA-molekyler. Vi valde RNA-molekyler vars uttryck är viktigt vid analys av bröstcancerceller, och optimerade metoden för analys av enskilda celler. Vi analyserade cancerceller från blodprov och använde metoden för att titta på RNA-nivåer i enskilda celler från en grupp av celler och visade på skillnader i RNA-nivåer inom gruppen. / <p>QC 20150115</p>

More than a decade after the completion of the human genome project, the function of a large number of genes remains to be elucidated. Forward and reverse genetic approaches have proven to be powerful tools to study gene function and have provided insights into fundamental biological processes. Furthermore, functional genetic screening can lead to a better understanding of the action of endogenous and exogenous stimuli such as hormones or drugs on biological systems. Thus far, systematic and unbiased studies have largely been limited to model organisms. However, complex disease-relevant genotypes and phenotypes cannot be studied in entirety in lower organisms creating a need for systematic approaches in human cells. This thesis describes a series of studies using forward and reverse genetic approaches combined with state-of-the-art technology in haploid human cells. The first chapter describes the development of a quantitative phenotypic read-out using a novel application of RNA-sequencing that allows the functional annotation of genes in signalling pathways. The presented data demonstrate that the employed shallow RNA-sequencing method is scalable and suitable as a read-out for reverse genetic screening. The second chapter focuses on the implementation of this method in a large reverse genetic study in human cells to functionally annotate tyrosine kinases in signalling pathways upon stimulation with a set of ten polypeptides and small molecules. The screens revealed known and unexpected interactions between different signalling molecules and pathways, validating the technical approach in a biological context. The third chapter presents a pilot study describing the set-up of a forward genetic technique for compound mode-of-action screening using a pooled human mutant cell line collection. The chemical genetic approach displayed sufficient sensitivity and allowed to monitor thousands of gene-drug interactions simultaneously. Together, this thesis combines elements to advance technological and biological aspects of functional genomics and chemical genetics.

Colonic stromal cells provide critical structural support but also regulate immunity, tolerance and inflammatory responses in the mucosa. Substantial variability and plasticity of mucosal stromal cells has been reported but a paucity of distinct marker genes exist to identify distinct cell states. Here single-cell RNA-sequencing is used to document heterogeneity and subtype specific markers of individual colonic stromal cells in human and mouse. Marker-free transcriptional clustering of fibroblast-like cells derived from healthy human tissue reveals distinct populations corresponding to myofibroblasts and three transcriptionally and functionally dissimilar populations of fibroblasts. A SOX6 high fibroblast subset occupies a position adjacent to the epithelial basement membrane and expresses multiple epithelial morphogens including WNT5A and BMP2. Additional fibroblast subtypes show specific enrichment for chemokine signalling and prostaglandin E₂ synthesis respectively. In ulcerative colitis, substantial remodelling occurs with depletion of the SOX6 high population and emergence of an immune enriched population expressing genes associated with fibroblastic reticular cells including CCL19, CCL21 and IL33. A large murine dataset comprising over 7,000 colonic mesenchymal cells from an acute colitis model and matched healthy controls reveals strong preservation of the SOX6 high and myofibroblast transcriptional signatures. Unsupervised pseudotemporal ordering is used to relate fibroblast subsets to one another producing a branched developmental hierarchy that includes a potential progenitor population with mesothelial characteristics at its origin. This work provides a molecular basis for re-classification of colonic stromal cells and identifies pathological changes in these cells underpinning inflammation in UC.

A teca é uma árvore de grande importância comercial pelas características de cor e durabilidade de sua madeira. Devido a sua rusticidade e fácil adaptação ao clima, plantios de teca tornam-se cada vez mais atrativos ao redor do mundo. Contudo, esta espécie apresenta escassez de estudos genéticos moleculares a respeito tanto de sua madeira, quanto de sua tolerância às variações ambientais. Uma vez que o transcriptoma pode apresentar grande quantidade de informação a respeito dos genes expressos por um conjunto celular, neste trabalho foi realizado o primeiro transcriptoma de teca, onde foram sequenciadas flores, folhas, raízes e seedlings pela tecnologia Illumina. A montagem do transcriptoma foi realizada com o programa Trinity acima de 100 milhões de reads e gerou mais de 400 mil contigs, os quais tiveram as anotações funcionais adquiridas com o programa Blas2GO. 51% dos contigs foram anotados, mostrando alta similaridade com as espécies Vitis vinifera e Solanum licopersicum; destes, 78% obtiveram anotações funcionais com o Gene Ontology, totalizando 5.165 termos para Processo Biológico, 2.846 termos para Função Molecular e 742 para Componente Celular. A expressão diferencial foi obtida com o programa edgeR a 5% de probabilidade de erro e mostrou que, para 187.315 contigs montados através da fusão de todas as bibliotecas sequenciadas, 18 mostraram expressão diferencial para flor, 14 para folha, 13 para raiz e 29 para seedling. Após a etapa de caracterização do transcriptoma, foi realizado um experimento de estresse por déficit hídrico em casa-de-vegetação, onde plantas de teca foram submetidas a estresse Moderado (40% de água no substrato por 20 dias), estresse Severo (20 a 40% de água por 30 dias) e tratamento controle (substrato saturado). As medições através de analisador de gases por infravermelho (IRGA) mostraram queda na fotossíntese (até 70% a menos do que o controle), na transpiração (até 77%) e na condutância estomática (até 85%) entre os tratamentos; além disto, o conteúdo relativo de água foliar caiu 13% entre o tratamento severo e o controle, e níveis de prolina livre foram até 3,5 vezes mais altos nos tratamentos de estresse. A temperatura foliar aumentou significativamente com o aumento da irradiância de fótons aplicada. A busca por genes relacionados ao estresse por déficit hídrico na biblioteca de transcritos de Raiz retornou 1.145 sequências, e destas, 4 foram caracterizadas: TgTPS (trealose 6-fosfato sintase), TgPIP (aquaporina, proteína intrínseca de membrana plasmática), TgDREB2 (proteína de ligação a elemento responsivo a desidratação) e TgAREB (proteína de ligação a elemento responsivo a ácido abscísico). Apenas TgTPS, TgPIP e TgDREB2 mostraram alto grau de conservação entre as espécies, podendo ser corretamente amplificadas via PCR e validadas por sequenciamento. Assim, com o banco de dados de transcritos obtido pelo RNA-seq, foi possível identificar genes candidatos ao estudo de características vegetativas e reprodutivas de teca, contribuindo para entender os mecanismos moleculares desta espécie florestal. / Teak is a tree of great commercial importance by the characteristics of color and durability of its wood. Due to its hardiness and easy adaptation to climate, teak plantations become increasingly attractive around the world. However, this species has a lack of molecular genetic studies on both of its wood, as their tolerance to environmental variations. Once the transcriptome can provide lots of information about the genes expressed by a cell group, this work represents the first transcriptome teak, which were sequenced flowers, leaves, roots and seedlings by Illumina technology. The transcriptome assembly was performed with Trinity program above 100 million reads and generated more than 400,000 contigs, which have acquired the functional annotations with Blas2GO program. 51% of the contigs were annotaded, showing high similarity to Vitis vinifera and Solanum licopersicum; of these, 78% had functional annotations with the Gene Ontology, totaling 5,165 terms for Biological Process, 2846 terms for Molecular Function and 742 for Cell Component. The differential expression was obtained with the edgeR program at 5% probability of error and showed that for 187,315 contigs assembled by merging all sequenced libraries, 18 showed differential expression to flower, 14 to leaf, 13 to root and 29 for seedling. After this step of characterization of the transcriptome, we performed a stress experiment by water deficit at greenhouse, where teak plants were subjected to Moderate stress (40% of water in the substrate for 20 days), Severe stress (20 to 40% water for 30 days) and control treatment (saturated substrate). Measurements by infrared gas analyzer (IRGA) showed a decrease in photosynthesis (up to 70% less than the control), transpiration (up 77%) and stomatal conductance (up 85%) between treatments; furthermore, leaf relative water content dropped 13% between the treatment control and severe, and free proline levels were up to 3.5 fold greater in stress treatments. The leaf temperature increased significantly with increasing irradiance of photons applied. The search for genes related to stress by water deficit in the root transcripts library returned 1,145 sequences, and these, 4 were characterized: TgTPS (trehalose 6-phosphate synthase), TgPIP (aquaporin, protein intrinsic of plasma membrane), TgDREB2 (dehydration responsive element binding protein) and TgAREB (abscisic acid responsive element binding protein). Only TgTPS, TgPIP and TgDREB2 showed a high degree of conservation between species, and can be properly amplified by PCR and validated by sequencing. Thus, with the database of transcripts obtained by RNA-seq, candidate genes were identified for the study of vegetative and reproductive characteristics teak, helping to understand the molecular mechanisms of this forest species.

The initiating events of ovarian cancer remain unknown, but an established risk factor is use of estrogen therapy by post-menopausal women where there is a positive correlation between duration of use and risk for disease. Mouse models of ovarian cancer have shown that exposure to exogenous 17β-estradiol (E2) accelerates tumour onset so this study aims to investigate the E2 signalling mechanisms responsible for sensitizing ovarian epithelial cells to transformation. By developing model systems that are responsive to E2 manipulation, we showed that E2 induces the formation of epithelial dysplasias both in vitro and in vivo. microRNA microarray was used to discover that E2 up-regulates microRNA-378 via the ESR1 pathway, resulting in the down-regulation of a tumour suppressor gene called Disabled-2 (Dab2). E2 suppression of Dab2 was found to result in increased proliferation, loss of contact inhibition, epithelial dysplasia, and increased sensitivity to transformation. This mechanism was also found to be active in mouse fallopian tube epithelium and human ovarian cancer cells. Single-cell RNA sequencing and trajectory analysis was subsequently used to explore additional signalling mechanisms that might contribute to the emergence of dysplastic lesions induced by E2. Multiple molecular signalling pathways dysregulated by E2 were identified and this revealed several possible biomarkers to be investigated for early detection of ovarian cancer. In the context of a current lack of strategies for ovarian cancer prevention or early detection, this work represents a significant advance in our understanding of how E2 promotes ovarian cancer initiation.