The Gram-negative bacterium Francisella tularensis is the causative agent of tularemia and a model intracellular pathogen. It is also considered a potential bioweapon, as F. tularensis is highly infectious and has the potential to cause fatal disease in humans. Many factors required for F. tularensis virulence have been identified, yet we know relatively little regarding how these factors are regulated at the level of transcription. In order to further understand the regulation of virulence factors in F. tularensis, we have systematically determined the genomic regions associated with all of the transcription factors implicated in virulence using chromatin immunoprecipitation coupled with high-throughput DNA sequencing (ChIP-Seq).

Somatic cells can be reprogrammed to induced pluripotent stem cells (iPSCs) by the ectopic expression of defined factors. iPSCs hold great promise for pharmaceutical screening and regenerative medicine but the mechanism of reprogramming is not well understood. This work examines a component process of reprogramming that is the mesenchymal-to-epithelial transition (MET), an important step in the generation of iPS cells. In this thesis I demonstrate a connection between BMP signaling and the reprogramming factor Klf4 in the activation of the MET expression program. Using ChIP-Seq I mapped the binding of Klf4 and BMP Smads across the genome and linked their co-binding to a MET expression program determined by RNA-Seq. My work uncovers a thus-far unreported interaction between Klf4 and BMP signaling in cellular epithelialization that can directly improve the technical methods of reprogramming and have important implications for the induction of epithelial tissues in general.

Somatic cells can be reprogrammed to induced pluripotent stem cells (iPSCs) by the ectopic expression of defined factors. iPSCs hold great promise for pharmaceutical screening and regenerative medicine but the mechanism of reprogramming is not well understood. This work examines a component process of reprogramming that is the mesenchymal-to-epithelial transition (MET), an important step in the generation of iPS cells. In this thesis I demonstrate a connection between BMP signaling and the reprogramming factor Klf4 in the activation of the MET expression program. Using ChIP-Seq I mapped the binding of Klf4 and BMP Smads across the genome and linked their co-binding to a MET expression program determined by RNA-Seq. My work uncovers a thus-far unreported interaction between Klf4 and BMP signaling in cellular epithelialization that can directly improve the technical methods of reprogramming and have important implications for the induction of epithelial tissues in general.

Multiple computational algorithms were developed for analyzing ChIP-seq datasets of histone modifications. For basic ChIP-seq data processing, the problems of ambiguous short sequence read mapping and broad peak calling of diffuse ChIP-seq signals were solved by novel statistical methods. Their performance was systematically evaluated compared with existing approaches. The potential utility of finding meaningful biological information was demonstrated by the applications on real datasets. For biological question driven data mining, several important topics were selected for algorithm developments, including hypothesis-driven insulator prediction, unbiased chromatin boundary element discovery and combinatorial histone modification signature inference. The integrative computational pipeline for insulator prediction not only produced a list of putative insulators but also recovered specific associated chromatin and functional features. Selected predictions have been experimentally validated. The unbiased chromatin boundary element prediction algorithm was feature-free and had the capability to discover novel types of boundary elements. The predictions found a set of chromatin features and provided the first report of tRNA-derived boundary elements in the human genome. The combinatorial chromatin signature algorithm employed chromatin profile alignments for unsupervised inferences of histone modification patterns. The signatures were associated with various regulatory elements and functional activities. Both the computational advantages and the biological discoveries were discussed.

Human autoimmune diseases are hallmarked by inappropriate loss-of-tolerance and self-attacking response of the immune system. Studies included in this thesis are focusing on the implication and functional impact of genetic factors in three autoimmune diseases rheumatoid arthritis (RA), asthma, and systemic lupus erythematosus (SLE). Using genetic association studies, we found in study I and II that sequence variants of the interferon regulatory factor 5 (IRF5) gene were associated with RA and asthma, and the associations were more pronounced in certain disease subtypes. Distinct association patterns or risk alleles of the IRF5 gene variants were revealed in different diseases, indicating that IRF5 contributes to disease manifestations in a dose-dependent manner. In study III, we found that seven out of eight genetic risk loci for SLE, which were originally identified in East Asian populations, also conferred disease risk with the same risk alleles and comparable magnitudes of effect sizes in Caucasians. Remarkable differences in risk allele frequencies were observed for all associated loci across ethnicities, which seems to be the major source of genetic heterogeneity for SLE. In study IV we explored an exhaustive spectrum of sequence variants in the genes inhibitor of kappa light polypeptide gene enhancer in B-cells kinase epsilon (IKBKE) and interferon induced with helicase C domain 1 (IFIH1) by gene resequencing, and identified nine variants in IKBKE and three variants in IFIH1 as genetic risk factors for SLE. One of the associated variants may influence splicing of IKBKE mRNA. In study V we provided genome-wide transcriptional regulatory profiles for IRF5 and signal transducer and activator of transcription 4 (STAT4) using chromatin immunoprecipitation-sequencing (ChIP-seq). The target genes of IRF5 and STAT4 were found to play active roles in pathways related with inflammatory response, and their expression patterns were characteristic for SLE patients. We also identified potential cooperative transcription factors for IRF5 and STAT4, and disease-associated sequence variants which may affect the regulatory function of IRF5 and STAT4. In conclusion, this thesis illuminates the contribution of several genetic risk factors to susceptibility of human autoimmune diseases, which facilitates our understanding of the genetic basis of their pathogenesis.

Mammalian cell types and tissues have diverse functional roles within an organism but can be derived by the differentiation of the embryonic stem cells (ESCs). ESCs are pluripotent cells with self-renewal properties. During development subsets of genes in ESCs are activated or silenced for manifestation of the cell type specific function. Gene expression changes occur transiently in early developmental stages, through signals received and executed by a variety of transcription factors (TFs), regulatory elements (promoters, enhancers) and epigenetic modifications of chromatin. Post-translational modifications of the histone tails are regulated by chromatin modifiers and transform the chromatin architecture. Polycomb (PcG) and Trithorax (TrxG) group proteins are the most commonly studied histone modifiers. They were first discovered as repressors (H3K27me3) and activators (H3K4me3) respectively of Homeobox (Hox) genes in Drosophila and they are conserved in mammals. Bivalent chromatin is defined as the simultaneous presence of silencing (H3K27me3) and activating (H3K4me3) histone marks and was first discovered as a feature of many developmental gene promoters of ESCs. Bivalent promoters are thought to be in a ‘poised’ state for later activation or repression during differentiation due to the presence of the two counter-acting histone modifications and a pausing variant of RNA polymerase II (RNAPII) accompanied with intermediate-low levels of expression. By integrative analysis of publicly available ChIP sequencing (ChIP-seq) datasets in murine and human ESCs, we predicted 3,659 and 4,979 high–confidence (HC) bivalent promoters in mouse and human ESCs respectively. Using a peak-based method, we acquire a set of bivalent promoters with high enrichment for developmental regulators. Over 85% of Polycomb targets were bivalent and their expression was particularly sensitive to TF perturbation. Moreover, murine HC bivalent promoters were occupied by both Polycomb repressive component classes (PRC1 and PRC2) and grouped into four distinct clusters with different biological functions. HC bivalent and active promoters were CpG rich while H3K27me3-only promoters lacked CpG islands. Binding enrichment of distinct sets of regulators distinguished bivalent from active promoters and a ‘TCCCC’ sequence motif was specifically enriched in bivalent promoters. Using the recent technology of single cell RNA sequencing (scRNA-seq) we focused on gene expression heterogeneity and how it may affect the output of differentiation. We collected single cell gene expression profiles for 32 human and 39 murine ESCs and studied the correlation between diverse characteristics such as network connectivity and coefficient of variation (CV) across single cells. We further characterized properties unique to genes with high CV. Highly expressed genes tended to have a low CV and were enriched for cell cycle genes. In contrast, High CV genes were co-expressed with other High CV genes, were enriched for bivalent promoters and showed enrichment for response to DNA damage and DNA repair. Bivalent promoters in ESCs grouped in four distinct classes of variable biological functions according to Polycomb occupancy and three RNAPII variants. To study the dynamics of epigenetic and transcription control at promoters during development, we collected ChIPseq data for two chromatin modifications (H3K4me3 and H3K27me3) and RNAPII (8WG16 antibody) as well as expression data (RNA-seq) across 8 cell types (ESCs and seven committed cell types) in mouse. Hierarchical clustering of 22,179 unique gene promoters across cell types, showed that H3K4me3 peaks are in agreement with the expression data while H3K27me3 and RNAPII peaks were not highly consistent with the hierarchical tree of gene expression. Unsupervised clustering of ChIP-seq and RNA-seq profiles has resulted in 31 distinct profiles, which were subsequently narrowed down to nine major profile groups across cell types. TF enrichment at individual clusters using ChIP sequencing data did not fully agree with the classification of 8 major profile groups. Considering all the above results, three major epigenetic profiles (active, bivalent and latent) seem to be conserved across the species and cell types in our study. These states could recapitulate only a fraction of the transcriptional information - adding other chromatin marks could enrich it - since they are seemingly unaffected by their respective expression profiles. H3K27me3 only state has low CpG density and shows stronger signatures at differentiated cell types. Transcriptional control is tighter in active than bivalent promoters and the different occupancy levels of PcG subunits and RNAPII can be reflected at the expression variance of bivalent genes, where a fraction of them are involved in developmental functions while others are more tissue-specific. Last, there is a striking similarity in the pausing patterns of RNAPII in the progenitor cell types, which suggests that RNAPII pausing is correlated with the developmental potential of the cell type. Finally, this analysis will serve as a resource for future studies to further understand transcriptional regulation during development.

We develop machine learning approaches to predict context specific enhancer-promoter interactions using evidence from changes in genomic protein occupancy over time. Occupancy of estrogen receptor alpha (ER-alpha), RNA polymerase (Pol II) and histone marks H2AZ and H3K4me3 were measured over time using ChIP-Seq experiments in MCF7 cells stimulated with estrogen. Two Bayesian classifiers were developed, unsupervised and supervised. The supervised approach uses the correlation of temporal binding patterns at enhancers and promoters and genomic proximity as features and predicts interactions. The method was trained using experimentally determined interactions from the same system and achieves much higher precision than predictions based on the genomic proximity of nearest ER-alpha binding. We use the method to identify a confident set of ER-alpha target genes and their regulatory enhancers genome-wide. Validation with publicly available GRO-Seq data shows our predicted targets are much more likely to show early nascent transcription than predictions based on genomic ER-alpha binding proximity alone. Accuracy of the predictions from the supervised model was compared against the second more complex unsupervised generative approach which uses proximity-based prior and temporal binding patterns at enhancers and promoters to infer protein-mediated regulatory complexes involving individual genes and their networks of multiple distant regulatory enhancers.

The motif discovery problem is becoming increasingly important for molecular biologists as new sequencing technologies are producing large amounts of data, at rates which are unprecedented. The solution space for DNA motifs is too large to search with naive methods, meaning there is a need for fast and accurate motif detection tools. We propose MotifGP, a multiobjective motif discovery tool evolving regular expressions that characterize overrepresented motifs in a given input dataset. This thesis describes and evaluates a multiobjective strongly typed genetic programming algorithm for the discovery of network expressions in DNA sequences. Using 13 realistic data sets, we compare the results of our tool, MotifGP, to that of DREME, a state-of-art program. MotifGP outperforms DREME when the motifs to be sought are long, and the specificity is distributed over the length of the motif. For shorter motifs, the performance of MotifGP compares favourably with the state-of-the-art method. Finally, we discuss the advantages of multi-objective optimization in the context of this specific motif discovery problem.

This thesis consists of two related studies directed at aspects of M. tuberculosis biology. The first focuses on deciphering gene-regulatory circuits from ChIP-seq data, and the second focuses on alternative strategies for combatting drug-resistant infections. The first study describes Binding Resolution Amplifier and Cooperative Interaction Locator (BRACIL), a post-peak-caller computational method that predicts transcription-factor (TF) binding sites with high-resolution as well as cooperative TF interactions derived from ChIP-seq data. BRACIL integrates ChIP-seq coverage with motif discovery from a signal-processing perspective and uses a blind-deconvolution algorithm that predicts binding-site locations and magnitudes. BRACIL also explicitly considers a second-order signal, represented by DNA fragments with two sites bound simultaneously, and uses it to predict cooperative interaction. Cooperative interaction indicates that the binding to a first site influences the probability of binding to a second site. This method estimates the probability of a binding configuration from the ChIP-seq coverage and performs a likelihood ratio test to predict cooperative interaction. As a proof of principle, I validated this method using M. tuberculosis transcription factor DosR. The second study focuses on strategies to fight antibiotic resistance. In particular, recent reports have shown the existence of treatment conditions (called "antiR") that select against drug-resistant strains. I used a mathematical model of infection dynamics and immunity to simulate the growth of resistant and sensitive pathogens under different treatment conditions (no drugs, antibiotic present, and antiR), and could show how a precisely timed combination of treatments can defeat resistant strains. This analysis suggested that a time- scheduled, multi-treatment therapy could lead to complete elimination of both sensitive and resistant strains. Also, my results indicated that the time necessary to turn a resistant infection into a sensitive one ("tclear") depends on the experimentally measurable rates of pathogen division, growth and plasmid loss. Additionally, I estimated tclear for a specific case, using available empirical data, and found that resistance may be lost up to 15 times faster under antiR treatment as compared to a no-treatment regime. Finally, an extension of these findings to population models provides quantitative support for therapeutic plans to clear antibiotic-resistant infections, including novel drug-cycling strategies.

Transcription factors activated by exogenous or endogenous stimuli alter gene expression with major effects on chromatin accessibility and the epigenome. This thesis investigates that impact of environmental chemical and hormonal exposure on liver chromatin accessibility in a mouse liver model. Exposure to the constitutive androstane receptor (CAR)-specific agonist ligand 1,4-bis-[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP) induces localized changes in chromatin accessibility at several thousand DNase hypersensitive sites (DHS). Activating histone marks, associated with enhancers and promoters, were induced by TCPOBOP and were highly enriched at opening DHS. Opening DHS were highly enriched for CAR binding sites and nuclear receptor direct repeat-4 motifs. These DHS were also enriched for the CAR heterodimeric partner RXRA, binding by CEBPA and CEBPB, and motifs for other liver-specific factors. Thus, TCPOBOP alters the enhancer landscape through changes in histone marks and by mechanisms linked to induced CAR binding. In other studies, the impact of pituitary growth hormone (GH) secretion patterns on chromatin accessibility changes associated with sex-biased liver gene expression was examined. In adult male liver, the transcription factor STAT5 is directly activated by each successive plasma GH pulse. In female liver, STAT5 is persistently activated by the near-continuous stimulation by plasma GH. A majority of the ~4,000 GH-regulated, sex-biased DHS have chromatin marks characteristic of enhancers and were enriched for proximity to sex-biased gene promoters. Chromatin accessibility is thus a key feature of sex-differential gene expression. Two major classes of male-biased DHS were identified: dynamic male-biased DHS, almost all bound by STAT5, which undergo repeated cycles of chromatin opening and closing induced by each GH pulse; and static male-biased DHS, whose accessibility is unaffected GH/STAT5 pulses and whose sex bias results from these chromatin sites being more closed in female liver. Sites with STAT5 binding showed greater chromatin opening, many of which also contain the STAT5 motif. Finally, the effect of a single GH pulse on hypophysectomized male mouse liver was investigated to identify DHS responsive to the male, pulsatile-GH, secretion pattern. These studies demonstrate that widespread epigenetic changes associated with target gene expression are induced by xenobiotics and hormones regulating liver gene expression. / 2022-01-31T00:00:00Z