Recent advances in global gene expression measurement and the development of large- scale public repositories for storage of such data have made a wealth of information available to researchers. While one gene expression study may lack sufficient replicates to make statistically significant pronouncements, the combination of studies through meta-analysis can yield results with a much greater likelihood of accuracy. In order to combine multiple sets of data, one must first address the issue of cross-comparison between global gene expression platforms, as well as resolve the issue of repeated measures (multiple probes representing the same gene) within each platform. In this work, I present computational methods for probe reannotation and scoring and for redundant probe consolidation that together allow for greatly improved access to data for meta-analysis. I also present an example of the application of these methods, in the analysis of the gene expression regulated by estrogen across multiple cell types. Estrogen, a steroid hormone, interacts with its receptors to regulate gene transcription in both direct and indirect manners. Estrogen has the effect of increasing proliferation in some tissues, while inhibiting proliferation or increasing apoptosis in others. How estrogen achieves these highly divergent results remains unclear. Through meta-analysis of gene expression experiments across multiple cell types, I show that patterns of estrogen regulation in many tissues involve the same key genes and pathways, including cell cycle, p53 signaling, and TGFβ signaling pathways. However, regulation in different cell types can result from regulation of different genes, or the same genes regulated in different directions. Many patterns of gene regulation support known physiological consequences of estrogen on these tissues. In particular, genes promoting proliferation are upregulated in uterus and certain breast and ovarian cancer cell lines. One gene, thrombospondin-1, is up-regulated in eleven out of nineteen cell types and may be a key player in regulating proliferation in re- sponse to estrogen. Results in other cell types are unexpected. Most notably, neither genes promoting nor inhibiting proliferation are differentially regulated upon estrogen treatment in vascular smooth muscle cells, despite estrogen inhibiting proliferation of these cells.

Transcription and translation are essential cellular processes that are coupled in bacteria. Though it was well-known that the rate of translation matches the rate of transcription, only in 2010 did evidence suggest direct physical coupling between the transcribing RNA polymerase (RNAP) and the translating ribosome. Nuclear magnetic resonance spectroscopy data showed that the RNAP-binding, transcription factor NusG could bind to the small ribosomal subunit protein, S10, through its C-terminal domain, thus, suggesting a model in which NusG simultaneously binds the transcription and translation machineries. In Chapter Two, I describe my investigations of the mechanism through which NusG-mediated transcription-translation coupling is established in bacteria, and how this coupling is regulated during gene expression. Specifically, I employed cell extract-based luciferase assays and purified C-terminal NusG mutants to show that the NusG N-terminal domain (NTD) and NusG F165A both inhibit transcription. This inhibitory effect was suppressed in an extract derived from a backtracking-resistant RNAP mutant strain, indicating that preventing backtracking by linking RNAP to the lead ribosome is a key function of NusG. While working with the cell extract-based luciferase assay system used to study NusG, I observed that deleting the cryptic rac prophage resulted in cell extracts with extremely low luciferase activity despite the strain having no phenotype in vivo. This initial observation grew into the project described in Chapter Three in which I explore the possibility of viral control of host genes by the poorly characterized rac prophage protein, RacR, through a combination of biochemical methods, structural modeling, bioinformatic analysis, and next-generation, transcriptome-wide, deep RNA sequencing. Taken together, the results reveal overlap between computationally predicted host gene targets and messenger RNA expression levels and suggest that RacR can function as a DNA-binding transcriptional regulator of host genes.

Maintaining cells in an undifferentiated totipotent state is essential for initiating developmental programs that lead to a fully formed organism in each generation and for maintaining immortal germ cells across generations. Caenorhabditis elegans is a powerful genetic model organism to study early germ cell development due to the animal’s transparency and the ability to screen for mutant phenotypes. However, our ability to use standard techniques to study gene expression using fluorescent reporter genes has been limited due to germline-specific silencing mechanisms that repress transgenes. Therefore, we lack even basic knowledge of how expression is regulated in C. elegans germ cells. In this study, we develop methods to overcome these silencing mechanisms by using a class of noncoding DNA, called Periodic An/Tn Clusters (PATCs), to prevent transgene silencing in the germline. We use these improved tools to test the proposed role of putative germline-specific regulatory DNA motifs and the role a periodic TT signal within germline promoters. We fused GFP to the promoter of a germline expressed gene (pcn-1), which is enriched for PATCs and contains a germline-specific motif (TTAAAG). Our results show that despite enrichment and phylogenetic conservation, the TTAAAG motif is not required for germline expression. To test additional motifs and periodic TTs, we have designed a system that will allow us to test synthetic gene fragments for bi-directional germline expression. These tools will allow us to rapidly test motif redundancy, motif spacing, and TT periodicity using gfp and rfp signals in the germline and will enable experiments aimed at understanding the role of germline regulatory elements.

A defining contribution of systems biology has been to reveal how cellular circuitry works to govern the state of a cell. Often, cell-state is determined by the activity of a small number of hyperconnected transcriptional regulators (TRs; e.g., transcription factors, (de)acetylases, (de)methylases, and other genes that act at the level of DNA to affect transcription). The activity of these TRs can be detected from the transcription of their targets, but doing so requires accurate gene regulatory networks (GRNs). The best way to construct GRNs is by combining computationally inferred networks with experimental perturbation data, but until recently this has not been feasible in human cells. As a first step in that direction, I undertook to perform a large-scale Transcriptional REgulator Knock-down (TREK), at two timepoints, in two cancer cell lines, at single-cell level, and to use the resulting data to improve our ability to infer the regulatory state of the cell. In all, I constructed regulons for over 900 TRs and described the dynamics both over time and across contexts. I have significantly improved our GRNs and, consequently, our ability to measure protein activity and identify cell-state regulators.

Roughly ~50% of the human genome, contains noncoding sequences serving as regulatory elements responsible for the diverse gene expression of the cells in the body. One very well studied category of regulatory elements is the category of enhancers. Enhancers increase the transcriptional output in cells through chromatin remodeling or recruitment of complexes of binding proteins. Identification of enhancer using computational techniques is an interesting area of research and up to now several approaches have been proposed. However, the current state-of-the-art methods face limitations since the function of enhancers is clarified, but their mechanism of function is not well understood. This PhD thesis presents a bioinformatics/computer science study that focuses on the problem of identifying enhancers in different human cells using computational techniques. The dissertation is decomposed into four main tasks that we present in different chapters. First, since many of the enhancer’s functions are not well understood, we study the basic biological models by which enhancers trigger transcriptional functions and we survey comprehensively over 30 bioinformatics approaches for identifying enhancers. Next, we elaborate more on the availability of enhancer data as produced by different enhancer identification methods and experimental procedures. In particular, we analyze advantages and disadvantages of existing solutions and we report obstacles that require further consideration. To mitigate these problems we developed the Database of Integrated Human Enhancers (DENdb), a centralized online repository that archives enhancer data from 16 ENCODE cell-lines. The integrated enhancer data are also combined with many other experimental data that can be used to interpret the enhancers content and generate a novel enhancer annotation that complements the existing integrative annotation proposed by the ENCODE consortium. Next, we propose the first deep-learning computational framework for identifying enhancers. The proposed system called Dragon Ensemble Enhancer Predictor (DEEP) is based on the novel deep learning two-layer ensemble algorithm capable of identifying enhancers characterized by different cellular conditions. Experimental results using data from ENCODE and FANTOM5, demonstrate that DEEP surpasses in terms of recognition performance the major systems for enhancer prediction and shows very good generalization capabilities in unknown cell-lines and tissues. Finally, we take a step further by developing a novel feature selection method suitable for defining a computational framework capable of analyzing the genomic content of enhancers and reporting cell-line specific predictive signatures.

Stomata are cellular valves in plants that play an essential role in the regulation of gas exchange and are distributed in the epidermis of aerial organs. In Arabidopsis thaliana, stomatal production and development are coordinated by the mitogen-activated protein kinase (MAPK) signalling pathway, which modulates a variety of other processes, including cell proliferation, regulation of cytokinesis, programed cell death, and response to abiotic and biotic stress. The environment also plays a role in stomatal development, by influencing the frequency at which stomata develop in leaves. This thesis presents an analysis of stomatal development in Arabidopsis mutants in two MAPK pathways: MEKK1-MKK1/MKK2-MPK4, and MAP3K17/18-MKK3. Obtained results demonstrate the effect of stress conditions on stomatal development and specify the involvement of analysed MAPK in stomatal patterning. First, both analysed pathways modulate stomatal patterning in Arabidopsis cotyledons. Second, plant growth-promoting bacteria tested enhance stomatal density and affect guard cell morphology. Third, the sucrose or mannitol treatment increases defects in stomatal patterning. Finally, salt stress or high temperature can suppress stomatal defects in mutants of the MEKK1-MKK1/MKK2-MPK4 pathway.

Mutations in genes that cause transcriptional dysregulation, such as genes that encode DNA and RNA-binding proteins (RNABPs), are a well-described cause of neurodevelopmental syndromes such as autism and epilepsy. Heterozygous de novo mutations involving the gene HNRNPU, which encodes the heterogeneous nuclear ribonuclear protein U, have been implicated in a neurodevelopmental syndrome most commonly characterized by epileptic encephalopathy. Although hnRNP U is a highly-abundant and ubiquitously-expressed DNA- and RNA-binding protein involved in a variety of important nuclear processes—most notably gene expression regulation—the role it plays in neurological disease is unclear and has yet to be studied. The work presented here examines a precision medicine approach for epilepsies thought to have a transcriptomic basis, starting with a thorough neurophysiological characterization of a heterozygous loss-of-function Hnrnpu mouse model (Hnrnpu+/113DEL), followed by a comprehensive and region-specific single-cell transcriptomic study, and finally the validation of implicated brain regions. Characterization of the Hnrnpu+/113DEL mouse line revealed an increased susceptibility to seizures in Hnrnpu+/113DEL mice, along with an increased perinatal mortality, global developmental delay and gait abnormalities. Gene expression profiling, including bulk RNA-sequencing of neocortex and single cell RNA-sequencing of both neocortex and hippocampus, revealed widespread, yet modest, dysregulation of gene expression that was largely inversely correlated to gene-length, and involved important, neurodevelopmental disease genes. In particular, pyramidal neurons of the subiculum displayed greater transcriptional burden upon heterozygous loss of Hnrnpu, with the known epilepsy gene Mef2c as a clear outlier showing greater than 50% reduction in expression. Follow-up investigation into whether this region- and cell-type specific gene dysregulation correlated to differences in neuronal function using c-Fos immunostaining, revealed an overall decrease in neuronal activity within the ventral subiculum in Hnrnpu+/113DEL mice. In summary, our data validates the presence of neurodevelopmental defects upon heterozygous loss of Hnrnpu and supports the notion of transcriptional dysregulation as a likely contributing factor to hnRNP U-related disease, possibly through the dysfunction of subiculum-derived excitatory neurons. Future studies evaluating the relationship between reduced activity within the ventral subiculum and hnRNP U disease phenotypes are an important next step, and may serve as the basis for targeted therapeutic discovery.

The development of Chromatin immunoprecipitation followed by sequencing (ChIP-seq) technology has enabled the construction of genome-wide maps of protein-DNA interaction. Such maps provide information about transcriptional regulation at the epigenetic level (histone modifications and histone variants) and at the level of transcription factor (TF) activity. This dissertation presents novel computational methods for ChIP-seq data analysis and applications. The work of this dissertation addresses four main challenges. First, I address the problem of detecting histone modifications from ChIP-seq cancer samples. The presence of copy number variations (CNVs) in cancer samples results in statistical biases that lead to inaccurate predictions when standard methods are used. To overcome this issue I developed HMCan, a specially designed algorithm to handle ChIP-seq cancer data by accounting for the presence of CNVs. When using ChIP-seq data from cancer cells, HMCan demonstrates unbiased and accurate predictions compared to the standard state of the art methods. Second, I address the problem of identifying changes in histone modifications between two ChIP-seq samples with different genetic backgrounds (for example cancer vs. normal). In addition to CNVs, different antibody efficiency between samples and presence of samples replicates are challenges for this problem. To overcome these issues, I developed the HMCan-diff algorithm as an extension to HMCan. HMCan-diff implements robust normalization methods to address the challenges listed above. HMCan-diff significantly outperforms another state of the art methods on data containing cancer samples. Third, I investigate and analyze predictions of different methods for enhancer prediction based on ChIP-seq data. The analysis shows that predictions generated by different methods are poorly overlapping. To overcome this issue, I developed DENdb, a database that integrates enhancer predictions from different methods. DENdb also integrates several experimental data including ChIP-seq data for TF binding sites. Finally, I present an extensive computational comparison of different ab-initio motif identification methods based on TF ChIP-seq data. The comparison included 10 different methods over 159 different TF datasets. Recommendations of this comparison indicate that the usage of simple methods outperforms the usage of high order models.

Respiratory syncytial virus (RSV) is the leading cause of respiratory illness in children worldwide. RSV has a negative sense RNA genome, which is the template for viral mRNA transcription and genome replication, and encodes a polymerase to carry out viral RNA synthesis. The promoters for RSV transcription and genome replication are found in a 44-nucleotide (nt), 3´-extragenic region called the leader (Le). Replication is initiated opposite the first nt of the Le, and transcription of the first gene begins at position +45, at a gene start (GS) sequence. However, transcription is also dependent on sequence within Le1-12. Interestingly, Le nucleotides 3-12 bear strong similarity to a GS signal. We hypothesized that this GS-like sequence is the recruitment site for transcribing polymerase. To test this hypothesis, we examined RNA synthesis events at the Le promoter. We identified a previously undescribed RNA initiation site at Le position +3 (Le+3) that was used frequently during RSV infection. Initiation at Le+3 led to the production of a small ~25 nt RNA. Le+3 initiation was shown to occur independently of replication initiation at +1, indicating it is a bona fide initiation site. Mutation of Le1-12 to increase similarity to a GS resulted in elongation of Le+3 RNA and a decrease in transcription initiation at the GS, demonstrating that the Le initiation sequence alters polymerase processivity and impacts downstream transcription events. Preliminary experiments to determine the function of the small RNA showed that it increased levels of viral RNA replication, suggesting it may be involved in influencing a switch from transcription to replication. These studies suggest a model for RSV transcription initiation, whereby the transcribing polymerase enters at the 3´–end of the genome, initiates RNA synthesis from Le+3 and generates a small RNA, and is then positioned to initiate transcription at the first GS. The small RNA that is generated may act as a feedback molecule to promote RNA replication. These findings provide a greater understanding of polymerase behavior at the promoter and may inform rational drug and vaccine design.

No description available.