Genregulation bezeichnet die geregelte Steuerung der Genexpression durch das Zusammenspiel einer Vielzahl von Transkriptionsfaktoren die in ihrer Gesamtheit hoch komplexe und zell-spezifische genregulatorische Netzwerke bilden. Im Rahmen meiner Arbeit beschäftigte ich mich mit zwei Ansätzen der computergestützten Analyse solcher Netzwerke, Modellierung und Reverse Engineering. Der erste Teil meiner Arbeit beschreibt die Entwicklung der Web-Anwendung GEne Network GEnerator (GeNGe). Hierbei handelt es sich um ein System für die automatische Erzeugung von genregulatorischen Netzwerken. Hierfür entwickelte und implementierte ich einen neuartigen Algorithmus für die Generierung von Netzwerkstrukturen die wichtige Eigenschaften biologischer Netzwerke zeigen. Für die dynamische Beschreibung der Transkription modifizierte ich eine nicht-lineare Kinetik. Diese neue Formulierung der Kinetik eignet sich besonders für die Erstellung von komplexen genregulatorischen Modellen am Computer. Desweiteren unterstützt GeNGe die Durchführung verschiedener in silico Experimente, um theoretische Aussagen über den Einfluss von Störungen des Systems treffen zu können. Der zweite Teil meiner Arbeit beschreibt die Entwicklung von GNRevealer. Es handelt sich hierbei um eine Methode zur Rekonstruktion von genregulatorischen Netzwerken auf Basis zeitdiskreter Messungen der Genexpression. Diese Methode verwendet ein neuronales Netz zusammen mit einem passenden Lernalgorithmus (backpropagation through time). Modifizierungen, welche notwendig für die Anwendung im Reverse Engineering Bereich sind, wurden von mir entwickelt, wie z.B. die Etablierung eines vollständigen Lernprozesses, die Diskretisierung der Ergebnisse und anschließende Validierungen. Im letzten Teil dieser Arbeit beschreibe ich eine Studie, in der sechs verschiedene Reverse Engineering Anwendungen von mir miteinander verglichen wurden. Diese Untersuchung hebt GNRevealer als geeignetste Anwendung aller getesteten Methoden hervor. / Gene regulation is accomplished mainly by the interplay of multiple transcription factors. This gives rise to highly complex and cell-type specific, interwoven structures of regulatory interactions summarized in gene regulatory networks. In this thesis, I address two approaches of computational analysis of such networks, forward modeling and reverse engineering. The first part of this thesis is about the Web application GEne Network GEnerator (GeNGe) which I have developed as a framework for automatic generation of gene regulatory network models. I have developed a novel algorithm for the generation of network structures featuring important biological properties. In order to model the transcriptional kinetics, I have modified an existing non-linear kinetic. This new kinetic is particularly useful for the computational set-up of complex gene regulatory models. GeNGe supports also the generation of various in silico experiments for predicting effects of perturbations as theoretical counterparts of biological experiments. Moreover, GeNGe facilitates especially the collection of benchmark data for evaluating reverse engineering methods. The second part of my thesis is about the development of GNRevealer, a method for reverse engineering of gene regulatory networks from temporal data. This computational approach uses a neural network together with a sophisticated learning algorithm (backpropagation through time). Specialized features developed in the course of my thesis include essential steps in reverse engineering processes such as the establishment of a learning workflow, discretization, and subsequent validation. Additionally, I have conducted a large comparative study using six different reverse engineering applications based on different mathematical backgrounds. The results of the comparative study highlight GNRevealer as best performing method among those under study.

Changes in cis- and trans-regulatory elements are among the prime sources of genetic and phenotypical variation at species level. The introduction of cis- and trans- regulatory variation has played important roles in driving diversity, phenotypical differentiation, and evolution of humans. Therefore, variation that occurs on cis- and trans- regulatory elements becomes imperative to better understanding of human genetic diversity and its evolution. In this research, around 3360 gene regulatory factors (GRF) from the human genome were catalogued. This catalog includes genes that code for proteins that perform gene regulatory activities such DNA-depending transcription, RNA polymerase II transcription cofactor and co-repressor activity, chromatin binding and remodeling, among other 218 regulatory functions. This GRF catalog allowed us to initially explore how some GRF genes have evolved in humans, archaic humans (Neandertal and Denisovan) and non-human primate species. We discussed the likely phenotypical and medical effects that evolutionary changes in GRF genes may have introduced into the human genome; for instance, traits associated to speech and language capabilities, genomic recombination hotspots, diseases, among others. By using genome-wide datasets, we additionally looked for GRFs likely to be candidates for positive selection in three human populations: Utah Residents with Northern and Western Ancestry (CEU), Han Chinese in Beijing (CHB), and Yoruba in Ibadan (YRI). As result, we produced a set of candidates that gathers genes that may have contributed in shaping the phenotypical diversity currently observed in these populations; for instance, by introducing regulatory diversity at population-specific level. We additionally identified six GRF classes enriched for genes located in regions that are likely candidates for positive selection at population specific level. We found that out of the 41 DNA-binding GRF classes classified so far, six groups exhibited enrichment for genes located on regions that may have been under positive selection: C2H2 zinc finger, KRAB-ZNF zinc finger, Homeo domain, Tryptophan cluster, Fork head/winged helix and, and High-mobility HMG domain. We additionally identified three KRAB-ZNF gene clusters, in the chromosomes one, three, and 16, for the Asian population that exhibit regions with extended haplotype homozygosity EHH (larger than 100 kb). This EHH suggests that these regions have undergone positive selection in CHB population. Finally, considering that a representative fraction of the phenotypic diversity observed between humans and its closely related species are likely explained by changes in cis-regulatory elements (CREs), we investigated putative binding sites for the transcription factor GABPa. Using ChIP-Seq data generated from a human cell line (HEK293T), 11,619 putative GABPa CREs were found, Out of which 224 are putative human-specific. To experimentally validate the transcriptional activity of these human-specific CREs, reporter gene essays and knock-down experiments were performed. Our results supported the functionality of these human-specific GABPa CREs and suggest that at least 1,215 genes are primary targets of GABPa. Finally, further analyses depict scenarios that put together transcriptional regulation by GABPa and the evolution of particular human traits; for instance, cognitive abilities, breast morphology, lipids and glucose metabolic pathways, among others.

Wie regulatorische Sequenzen die Genexpression steuern, ist von grundlegender Bedeutung für die Erklärung von Phänotypen in Gesundheit und Krankheit. Die Funktion regulatorischer Sequenzen muss letztlich in ihrer genomischen Umgebung und in entwicklungs- oder gewebespezifischen Zusammenhängen verstanden werden. Da dies eine technische Herausforderung ist, wurden bisher nur wenige regulatorische Elemente in vivo charakterisiert. Hier verwenden wir Induktion von Cas9 und multiplexed-sgRNAs, um hunderte von Mutationen in Enhancern/Promotoren und 3′ UTRs von 16 Genen in C. elegans zu erzeugen. Wir quantifizieren die Auswirkungen von Mutationen auf Genexpression und Physiologie durch gezielte RNA- und DNA-Sequenzierung. Bei der Anwendung unseres Ansatzes auf den 3′ UTR von lin-41, bei der wir hunderte von Mutanten erzeugen, stellen wir fest, dass die beiden benachbarten Bindungsstellen für die miRNA let-7 die lin-41-Expression größtenteils unabhängig voneinander regulieren können, mit Hinweisen auf eine mögliche kompensatorische Interaktion. Schließlich verbinden wir regulatorische Genotypen mit phänotypischen Merkmalen für mehrere Gene. Unser Ansatz ermöglicht die parallele Analyse von genregulatorischen Sequenzen direkt in Tieren. / How regulatory sequences control gene expression is fundamental for explaining phenotypes in health and disease. The function of regulatory sequences must ultimately be understood within their genomic environment and development- or tissue-specific contexts. Because this is technically challenging, few regulatory elements have been characterized in vivo. Here, we use inducible Cas9 and multiplexed guide RNAs to create hundreds of mutations in enhancers/promoters and 3′ UTRs of 16 genes in C. elegans. We quantify the impact of mutations on expression and physiology by targeted RNA sequencing and DNA sampling. When applying our approach to the lin-41 3′ UTR, generating hundreds of mutants, we find that the two adjacent binding sites for the miRNA let-7 can regulate lin-41 expression largely independently of each other, with indications of a compensatory interaction. Finally, we map regulatory genotypes to phenotypic traits for several genes. Our approach enables parallel analysis of gene regulatory sequences directly in animals.

Changes in cis- and trans-regulatory elements are among the prime sources of genetic and phenotypical variation at species level. The introduction of cis- and trans- regulatory variation has played important roles in driving diversity, phenotypical differentiation, and evolution of humans. Therefore, variation that occurs on cis- and trans- regulatory elements becomes imperative to better understanding of human genetic diversity and its evolution. In this research, around 3360 gene regulatory factors (GRF) from the human genome were catalogued. This catalog includes genes that code for proteins that perform gene regulatory activities such DNA-depending transcription, RNA polymerase II transcription cofactor and co-repressor activity, chromatin binding and remodeling, among other 218 regulatory functions. This GRF catalog allowed us to initially explore how some GRF genes have evolved in humans, archaic humans (Neandertal and Denisovan) and non-human primate species. We discussed the likely phenotypical and medical effects that evolutionary changes in GRF genes may have introduced into the human genome; for instance, traits associated to speech and language capabilities, genomic recombination hotspots, diseases, among others. By using genome-wide datasets, we additionally looked for GRFs likely to be candidates for positive selection in three human populations: Utah Residents with Northern and Western Ancestry (CEU), Han Chinese in Beijing (CHB), and Yoruba in Ibadan (YRI). As result, we produced a set of candidates that gathers genes that may have contributed in shaping the phenotypical diversity currently observed in these populations; for instance, by introducing regulatory diversity at population-specific level. We additionally identified six GRF classes enriched for genes located in regions that are likely candidates for positive selection at population specific level. We found that out of the 41 DNA-binding GRF classes classified so far, six groups exhibited enrichment for genes located on regions that may have been under positive selection: C2H2 zinc finger, KRAB-ZNF zinc finger, Homeo domain, Tryptophan cluster, Fork head/winged helix and, and High-mobility HMG domain. We additionally identified three KRAB-ZNF gene clusters, in the chromosomes one, three, and 16, for the Asian population that exhibit regions with extended haplotype homozygosity EHH (larger than 100 kb). This EHH suggests that these regions have undergone positive selection in CHB population. Finally, considering that a representative fraction of the phenotypic diversity observed between humans and its closely related species are likely explained by changes in cis-regulatory elements (CREs), we investigated putative binding sites for the transcription factor GABPa. Using ChIP-Seq data generated from a human cell line (HEK293T), 11,619 putative GABPa CREs were found, Out of which 224 are putative human-specific. To experimentally validate the transcriptional activity of these human-specific CREs, reporter gene essays and knock-down experiments were performed. Our results supported the functionality of these human-specific GABPa CREs and suggest that at least 1,215 genes are primary targets of GABPa. Finally, further analyses depict scenarios that put together transcriptional regulation by GABPa and the evolution of particular human traits; for instance, cognitive abilities, breast morphology, lipids and glucose metabolic pathways, among others.

X-Chromosom-Inaktivierung (XCI) ist der Mechanismus, den Säuger zur Dosiskompensierung zwischen weiblichen und männlichen Zellen verwenden. XCI wird ausgelöst durch die monoallelische Hochregulation der langen nicht-kodierenden RNA Xist von einem der zwei X-Chromosomen in weiblichen Zellen. Die Xist RNA vermittelt dann das Ausschalten der Gene auf diesem X-Chromosom. Das wirft einige interessante Fragen auf: Wie zählen Zellen ihre X-Chromosomen und stellen sicher, dass genau eines aktiv bleibt? Wie entscheiden sie, welches X-Chromosom aktiv bleibt und welches ausgeschaltet wird? Und wie erinnern sie sich an diese Entscheidung und behalten sie stabil bei durch alle weiteren Zellteilungen? Mithilfe eines stochastischen Modells zeigen wir, dass diese XCI Regulation prinzipiell durch nur zwei Regulatoren erklärt werden kann: Ein global (in trans) agierender XCI Aktivator und ein lokal (in cis) agierender XCI Repressor. Dieses Netzwerk aus nur zwei Regulatoren kann die Xist Expressionsmuster in verschiedenen Säugerspezies reproduzieren, von der Maus bis zum Mensch. Es sagt außerdem voraus, dass Zellen in der Lage sind, biallelische zu monoallelischer Xist Expression zu korrigieren, eine Vorhersage, für die wir tatsächlich experimentelle Belege finden. Mit einem mechanistischen Modell zeigen wir, dass das cis-Gedächtnis über den Xist Expressionszustand durch Antisense-Transkription zustande kommen könnte. Auf dieser Hypothese aufbauend untersucht der zweite Teil der Arbeit das Potential von Antisense-Transkription, ein lokales Gedächtnis über den Expressionszustand eines Gens zu generieren, genauer. Diese Analyse sagt vorher, dass Antisense-Repression den Expressionszustand eines Lokus tatsächlich für einige Tage stabil erhalten kann. / X-chromosome inactivation (XCI) is the mechanism for dosage compensation between the sexes in mammals. It is initiated through monoallelic upregulation of the long non-coding RNA Xist from one X chromosome, which mediates almost complete transcriptional silencing of this X chromosome. XCI regulation raises intriguing and thus far unanswered questions: How do cells count their X chromosomes and ensure that exactly one stays active? How do they make a mutually exclusive choice for one inactive X chromosome, and how do they then stably maintain this choice throughout subsequent cell divisions? Using stochastic modeling, we show that XCI onset only requires two regulators: A trans-acting Xist activator that ensures female specificity and a cis-acting Xist repressor that allows stable maintenance of alternative Xist expression states. This two-regulator network can recapitulate Xist expression patterns across different species and makes a novel prediction that is validated experimentally: Cells are able to revert biallelic Xist expression to monoallelic expression. With a mechanistic stochastic model we show that Xist's antisense transcript Tsix might be the cis-acting Xist repressor, uncovering the molecular mechanism behind the stabilization of the alternative Xist expression states. Building upon Tsix' possible functional role in stabilizing alternative Xist expression states on the active and inactive X chromosome, the second part of this thesis investigates the potential of antisense transcription to maintain a transient transcriptional memory. We find that mutual repression between a pair of antisense genes can allow the locus to remember the transcription state it has acquired due to a past signal for several days.