Genomic imprinting is the epigenetic modification of loci, primarily by DNA methylation, which results in parent-of-origin-specific monoallelic expression of a small subset of genes. In plants, imprinting occurs during endosperm development and a balance of maternally- and paternally-expressed imprinted genes is essential for normal seed development. Dependence on DNA methylation for imprinting highlights the potential to manipulate seed development, and consequently seed size, by altering DNA methyltransferase activity. DNA METHYLTRANSFERASE 1 (MET1) is the primary plant maintenance DNA methyltransferase and plays a significant role in imprinting. However, no evaluation of the potential role for other MET1 family members in genomic imprinting has been reported. The current model for the control of imprinting in plants suggests that maintenance DNA methyltransferases are required throughout development, yet the tissue-specific requirement of these enzymes is unconfirmed as analysis has relied solely on constitutive DNA methyltransferase mutants. To address these problems and to evaluate the potential to alter seed size, the work reported in this thesis investigated the potential involvement of putative maintenance DNA methyltransferases MET2a, MET2b and MET3 and the tissue-specific role of MET1 in imprinting. Imprinting was not significantly altered in met2a-1, met2b-1 and met3-1 mutants, indicating that MET1 is the sole DNA methyltransferase required for imprinting. Transcriptional analysis suggested MET1 is expressed throughout floral organ development and in the male and female gametophyte generation indicating that MET1 is potentially available to maintain imprinting-dependent methylation in these tissues. Tools to suppress MET1 tissuespecifically were developed to investigate the tissue-specific requirement of MET1 for imprinting. Analysis indicates that such tools could also be used to alter seed size by manipulating imprinting in commercially important species. Further work is needed to validate this approach.

Epigenetic states constitute heritable features of the chromatin to regulate when, where and how genes are expressed in the developing conceptus. A special case of epigenetic regulation, genomic imprinting, is defined as parent of origin-dependent monoallelic expression. The Igf2-H19 locus is considered as paradigm of genomic imprinting with a growth-promoting gene, Igf2, expressed paternally and a growth antagonist, H19 encoding a non-coding transcript, expressed only from the maternal allele. The monoallelic expression patterns are regulated by the epigenetic status at an imprinting control region (ICR) in the 5´-flank of the H19 gene. The chromatin insulator protein CTCF interacts with only the maternal H19 ICR allele to prevent downstream enhancers to communicate with the Igf2 promoters. Mutations of these CTCF binding sites lead to biallelic Igf2 expression, increased size of the conceptus and predisposition for cancer. Reasoning that these effects cannot be explained by the regulation of Igf2 expression alone, a technique was invented to examine long-range chromatin interactions without prior knowledge of the interacting partners. Applying the circular chromosomal conformation capture (4C) technique to mouse neonatal liver cells, it was observed that 114 unique sequences interacted with the H19 ICR. A majority of these interactors was in complex with only the maternal H19 ICR allele and depended on the presence of functional CTCF binding sites. The functional consequence of chromosomal networks was demonstrated by the observation that the maternal H19 ICR allele regulated the transcription of two genes on another chromosome. As the chromosomal networks underwent reprogramming during the maturation of embryonic stem cells, attention was turned to human cancer cells, displaying features common with mouse embryonic stem cells. Subsequently, chromatin folding at the human H19 ICR suggested that stable chromatin loops were organized by synergistic interactions within and between baits and interactors. The presence of these interactions was linked to DNA methylation patterns involving repeat elements. A "flower" model of chromatin networks was formulated to explain these observations. This thesis has unravealed a novel feature of the epigenome and its functions to regulate gene expression in trans. The identified roles for CTCF as an architectural factor in the organization of higher order chromatin conformations may be of importance in understanding development and disease ontogeny from novel perspectives.

Les síndromes de Prader-Willi (SPW) i d'Angelman (SA) són causades per diferents anomalies genètiques que afecten la regió 15q11-q13 regulada per impressió genètica. La manca d'expressió de gens paterns causa la SPW i anomalies que afecten la còpia materna del gen UBE3A causa la SA. El 70-75% del casos SPW i SA s'originen per delecions de la regió cromosòmica 15q11-q13 en el cromosoma patern o matern respectivament. Un 20-25% dels casos SPW són deguts a disomies uniparentals (DUP) maternes i el 2-5% dels casos SA són causats per DUP paternes. Entre l'1 i el 5% dels casos són causats per un defecte en la impressió (DI) en ambdues síndromes. En la SA la segona causa més freqüent (10-15%) són mutacions puntuals en el gen UBE3A i un percentatge de casos similar a l'anterior presenten una clínica consistent de la SA però es desconeix la causa genètica. Les delecions i DUP tenen un risc de recurrència molt baix. Aproximadament el 85% del casos de DI no presenten delecions ni anomalies en la seqüència del centre regulador de la impressió genètica (CI) i són considerats errors epigenètics amb un risc de recurrència molt baix. En el 15% restant el DI és originat per una deleció en el CI que pot ser familiar comportant un risc del 50% o bé pot ser de novo.L'objectiu principal de la present tesi doctoral va ser millorar les tècniques diagnòstiques de la SPW i la SA i aprofundir en la etiologia d'ambdues síndromes. Per tal d'assolir aquest objectiu es va realitzar la posta a punt de la tècnica M-PCR, per analitzar el patró de metilació del CI que permet el diagnòstic de les síndromes quan la causa és una deleció de la regió 15q11-q13, o una DUP o un DI. Tanmateix s'ha analitzat la seqüència del gen UBE3A en 30 pacients SA amb patró de metilació normal i clínica consistent, i s'ha realitzat l'estudi del CI en els casos SPW i SA amb un DI. Com resultat d'aquests estudis s'han identificat cinc mutacions en el gen UBE3A i una deleció de novo en el centre d'impressió en un pacient SPW. De les cinc mutacions en el gen UBE3A tres han estat de novo i dues familiars. A més d'aquestes cinc mutacions tres no havien estat descrites amb anterioritat. En la resta de pacients amb DI s'ha descartat la presència de deleció i de mutacions puntuals pel que el defecte d'impressió ha estat causat per un error epigenètic.Aquests resultats ens han permès l'assessorament genètic als familiars de risc i oferir un diagnòstic prenatal així com valorar les correlacions fenotip-genotip en la SPW i SA. / Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are caused by different genetic abnormalities affecting the 15q11-q13 region regulated by imprinting. Loss of paternal expression genes causes PWS and genetic abnormalities in the maternal UBE3A gene cause AS. Paternal or maternal 15q11-q13 deletions cause 70-75% of PWS and AS cases, respectively. Maternal uniparental disomy (UPD) cause 20-25% of PWS cases and 2-5% of AS cases are caused by paternal UPD. The range 1-5% of cases is due to imprinting defects (ID) in both syndromes. The second cause of AS (10-15%) is an UBE3A gene mutation and in the same percentage of AS patients the genetic cause is unknown. Deletions and UPD have a low recurrence risk. The 85% of ID do not have abnormalities in the imprinting center (IC) and are considered epigenetic defects with a low recurrence risk. The 15% of ID are caused by an IC deletion that could be hereditary with a 50% of recurrence risk or could be sporadic.The aim of this thesis was to improve the PWS and AS diagnostic techniques and to study in depth the etiology of both syndromes. In order to achieve this aim we optimized the M-PCR technique to analyze the IC methylation patern for 15q11-q13 deletions, UPD and ID diagnosis. Moreover, UBE3A sequence analysis was done in 30 AS patients with normal methylation patern and IC study was done in 2 PWS and 4 AS cases with an ID. As a result of these studies we identified five UBE3A mutations and a de novo IC deletion in a PWS ID patient. Out of the five UBE3A mutations three were sporadic and two hereditary. Moreover, three of these five UBE3A mutations have not been previously described. In the other ID patients we conclude the cause is an epimutation since the presence of a deletion or mutations in the IC was rule out.These results have allowed to offer recurrence risk and prenatal diagnosis and to assess phenotype-genotype correlations.

Angelman syndrome (AS) is a rare genetic disorder affecting 1/10,000 to 1/20,000 births. This disorder arises through the genetic disruption of the maternal UBE3A allele, which when coupled with epigenetic silencing of the paternal allele UBE3A allele, gives rise to an absence of UBE3A protein in the central nervous system. Clinical manifestations of the syndrome vary in severity and include poor motor function, deficits in language and severe intellectual impairments. Previous research in the Angelman syndrome mouse model revealed abnormalities in dendritic spine density and morphology of hippocampal pyramidal cells. As seen in humans with AS, mice show abnormal behavioral characteristics that include deficits in motor coordination and ability as well as hippocampal dependent associative fear conditioning. Physiologically, these animals exhibit severe deficits in long-term potentiation (LTP) when compared to wildtype littermates. In an attempt to reduce the time from laboratory study to human translation, we began to search a small molecule library for established compounds with the ability to improve the behavioral and physiological defects normally associated with the AS mouse. One compound, minocycline, was found to normalize the density of dendritic spines in the hippocampus as well as recover the associative memory of AS mice. Moreover, a significant increase in LTP after theta-burst stimulation was also observed in area CA1 hippocampal pyramidal neurons of AS mice treated with minocycline when compared to saline vehicle control mice. These results suggest treatment with minocycline improves synaptic function and learning and memory of AS mice and may provide similar improvements to patients with Angelman syndrome. Twenty-five participants ages 4-12 were enrolled in a clinical trial examining the safety and tolerability of minocycline in children with Angelman syndrome. Patients were evaluated at 3 time points, baseline (T1), after 8 weeks of treatment (T2) and again 8 weeks after the drug was discontinued (T3). Each evaluation was identical and included laboratory testing, EEG recording and neuropsychological examination. Results of the study showed minocycline was safe and well tolerated with only minor adverse effects reported. While no change was observed in EEG recordings, significant increases in the mean clinical global impressions severity scale score were observed after treatment with minocycline. Moreover, participants showed significant improvement in the raw scores of the communication and self-care domains of the Bayley Scale of Infant and Toddler Development, 3rd Edition. These results show for the first time, a therapeutic with the ability to improve the characteristic behaviors of Angelman syndrome. Unfortunately, currently available neuropsychological measures were found to be insensitive to the behavioral phenotype of AS. The primary outcome measure, the Bayley Scale of Infant and Toddler Development, 3rd Edition relies on verbal communication and for the examinee to perform specific tasks on command. These testing methods are not compatible with this patient population and resulted in raw scores outside of 2 standard deviations from the mean. The inability of the participants to perform on these exams led us to develop a novel outcome measure; one that relies on observation rather than verbal communication. 9 children with AS and 7 healthy children were enrolled in an observational study in which 30 minutes of free play activity was video recorded. Using behavioral coding software, 3 independent raters quantified stereotypical AS behaviors as well as communication methods. Speech attempts were categorized into five difference types of vocalizations and revealed children with AS used less advance forms of vocalization consisting mostly of phonation control. Phonetic inventories show mostly front or back vowel usage suggesting little tongue movement occurs during speech production. These results suggest deficits in verbal ability may be related to a childhood apraxia of speech. Impairments in balance and motor coordination have been associated with AS. In an attempt to measure gross motor function, spatiotemporal gait parameters were collected using an electronic walkway and gait analysis software. Results show the gait of children with AS most resembles that of patients with ataxia but without cognitive impairment. In an attempt to develop a single quantitative measure able to describe the severity of gait-related disability, statistical methods were used to create a gait index for patients with AS. The results of this study provides AS researchers with the tools necessary to accurately measure changes in behavior and gait during the clinical evaluation of potential therapeutics

published_or_final_version / Statistics and Actuarial Science / Master / Master of Philosophy

Genetic association study is a useful tool to identify the genetic component that is responsible for a disease. The phenomenon that a certain gene expresses in a parent-of-origin manner is referred to as genomic imprinting. When a gene is imprinted, the performance of the disease-association study will be affected. This thesis presents statistical testing methods developed specially for nuclear family data centering around the genetic association studies incorporating imprinting effects. For qualitative diseases with binary outcomes, a class of TDTI* type tests was proposed in a general two-stage framework, where the imprinting effects were examined prior to association testing. On quantitative trait loci, a class of Q-TDTI(c) type tests and another class of Q-MAX(c) type tests were proposed. The proposed testing methods flexibly accommodate families with missing parental genotype and with multiple siblings. The performance of all the methods was verified by simulation studies. It was found that the proposed methods improve the testing power for detecting association in the presence of imprinting. The class of TDTI* tests was applied to a rheumatoid arthritis study data. Also, the class of Q-TDTI(c) tests was applied to analyze the Framingham Heart Study data. The human microbiome is the collection of the microbiota, together with their genomes and their habitats throughout the human body. The human microbiome comprises an inalienable part of our genetic landscape and contributes to our metabolic features. Also, current studies have suggested the variety of human microbiome in human diseases. With the high-throughput DNA sequencing, the human microbiome composition can be characterized based on bacterial taxa relative abundance and the phylogenetic constraint. Such taxa data are often high-dimensional overdispersed and contain excessive number of zeros. Taking into account of these characteristics in taxa data, this thesis presents statistical methods to identify associations between covariate/outcome and the human microbiome composition. To assess environmental/biological covariate effect to microbiome composition, an additive logistic normal multinomial regression model was proposed and a group l1 penalized likelihood estimation method was further developed to facilitate selection of covariates and estimation of parameters. To identify microbiome components associated with biological/clinical outcomes, a Bayesian hierarchical regression model with spike and slab prior for variable selection was proposed and a Markov chain Monte Carlo algorithm that combines stochastic variable selection procedure and random walk metropolis-hasting steps was developed for model estimation. Both of the methods were illustrated using simulations as well as a real human gut microbiome dataset from The Penn Gut Microbiome Project. / published_or_final_version / Statistics and Actuarial Science / Doctoral / Doctor of Philosophy

Traditionally, autosomal genes, when expressed, are assumed to express both alleles equally. Exceptions to this tenet include genes for which a specific genotypic polymorphism controls expression level, as well as genes for which monoallelic expression is epigenetically dictated. In this work, we discovered and characterized allele-specific gene expression in a variety of tissues using multiple techniques. We used cell lines and fresh tissue from reciprocal crosses of F1 heterozygous mice and the homozygous parental strains. We relied on a variety of high-throughput genomic techniques including RNA-Seq and DNA SNP-arrays to examine multiple types of allele-specific expression. We searched for novel examples of random autosomal monoallelic expression (RMAE) by using DNA SNP-arrays and cDNA from lymphoblast and fibroblast clonal cell lines of heterozygous mice. We found that of the approximately 1,350 autosomal genes we assessed, over 10% showed evidence of RMAE in at least one cell type. This allele-specific expression was stable over long periods in cell culture and encompassed a variety of gene types, some of which also exhibit RMAE in human clonal lines. Additionally, for a subset of RMAE genes, there seemed to be preferential inactivation of one allele; this monoallelic expression was still considered random, as from clone to clone the gene could be expressed either monoallelically or biallelically. In a second set of experiments we developed an analysis pipeline to examine RNASeq data for allele-specific expression. Using a pilot data set, we assessed the murine liver for parent-of-origin monoallelic expression, examining both known and candidate novel imprinted genes. We observed imprinted monoallelic expression in the adult liver for some, but not all, imprinted genes that have been reported in studies of embryonic tissue. The results suggest that there are few, if any, novel imprinted genes to be discovered in the mouse liver. This pilot data set also allowed examination of the genetic basis of allele specific gene expression. In keeping with recent reports, we found evidence for genetically based allele-specific expression, ranging from mild to greater than 4-fold allelic imbalance. We examined the extent to which this allelic imbalance correlated with total expression levels in the parental strains.

Genomic imprinting, the parent-of-origin specific silencing of alleles, plays an important role in phenotypic plasticity and consequently evolution. The leading explanation for genomic imprinting is Haig's conflict theory, which suggests that alleles from each parent have evolved under different selectional pressures, resulting in the differential expression of patrigenes and matrigenes. Previous studies have mainly used mammals and flowering plants to test Haig’s theory. However, there is a lack of independent evidence to support the theory. My PhD thesis attempts to conduct an independent test of Haig’s conflict theory using buff tailed bumblebee Bombus terrestris. A methylation system to facilitate genomic imprinting has not been found in this species. Therefore the first aim of the study was to establish the presence of a functional methylation system in B. terrestris before testing Haig's conflict theory using worker reproduction in queen-less colonies. The initial finding is that a methylation system exists in B. terrestris. The next study, investigating the presence of methylated genes, revealed differential methylation patterns in caste and life stages. Finally, genes involved with worker reproduction in a range of social insects were identified, but distinguishing the matrigene and the patrigene for each gene was unsuccessful. Therefore the final study investigating the presence of imprinted genes in B. terrestris and whether they conform to the expression patterns hypothesised by Haig’s conflict theory could not be analysed. Although this study did not provide conclusive evidence to support Haig’s conflict theory, the presence of methylation in genes involved with worker reproduction in reproducing and non-reproducing B. terrestris workers suggests that further analysis is needed. With adequate evidence, proving Haig’s conflict theory will not only expand our knowledge of invertebrate methylation, but also our understanding of conflict within social insect societies and our knowledge of how genomic imprinting affects phenotypic plasticity.

Transgenic organisms are a valuable tool for studying epigenetics, as they provide significant insight into the evolutionary conservation of epigenetic control sequences, the interacting proteins, and the underlying molecular mechanisms. Paramutation is an epigenetic phenomenon in which the epigenetic status and expression level of one allele is heritably altered after pairing with another. At the b1 locus in maize, a control region consisting of seven 853 bp tandem repeats is required for paramutation. To study the conservation of the epigenetic mechanisms underlying maize b1 paramutation, I created transgenic Drosophila carrying the maize b1 control region flanked by FRT sites and adjacent to the Drosophila white reporter gene. The maize b1 tandem repeats caused epigenetic silencing in Drosophila, as white expression consistently increased following repeat removal. A single copy of the tandem repeat sequence was sufficient to cause silencing, and silencing strength increased as the number of repeats increased. Trans interactions, such as pairing-sensitive silencing, were also observed and appear to require a threshold number of b1 tandem repeats, similar to paramutation in maize. Analysis of transcription from the repeats showed that the b1 tandem repeats are transcribed from both strands in Drosophila, as they are in maize. Bidirectional transcription was found to extend to the regions flanking the repeats, and persisted in “repeats-out” transgenes following repeat removal. However, aberrant transcription was lost when a zero-repeat transgene was moved to a new genomic position, suggesting that it may be due to an epigenetic mark that is retained from the previous silenced state. A search for modifiers of b1 repeat-mediated silencing demonstrated that Polycomb group proteins are involved. Together, these results indicate considerable conservation of an epigenetic silencing process between the plant and animal kingdoms. Genomic imprinting is a related epigenetic process in which parent-specific epigenetic states are inherited and maintained in progeny. The conservation of epigenetic mechanisms was further explored via an in-depth review of the molecular mechanisms underlying genomic imprinting in plants, mammals and insects, and identification of potentially imprinted genes in Drosophila by microarray analysis.

Surface modification is an extensively researched approach in order to overcomethe limitations, and improve the performance of orthopedic and dental implants. It is atthe surface of the implant materials that the initial interactions of tissues or body fluidstake place. Therefore, surface properties of biomaterials are the important factors that cancontrol these biological responses. Molecular imprinting is a surface modificationtechnique that creates specific recognition sites on the surface of biomaterials. Todevelop the recognition sites, a functional monomer is assembled with templatebiomolecule and then crosslinked. After removal of the template, the surface can rebindthe molecules. Therefore, desired reactions can be initiated at the interface between tissueand implants by modifying surfaces to selectively bind certain types of biomolecules,such as proteins. The objective of this project was to observe the potential of molecularimprinting technique for creating biomaterials that can recognize specific biomolecules.Fluorescently labeled lysozyme or RNase A was used as a template biomolecule and theprotein-imprinted scaffolds were fabricated by sol-gel processing. To interpret the densityof binding sites created, the quantity of surface-accessible protein was determined. Theamount of protein available on the surface was proportional to the amount loaded.Protein-imprinted scaffolds were evaluated for their ability to selectively recognize thetemplate biomolecule. Further, for these selectivity studies, a combination of theimprinted protein and a competitor protein were rebound to the polysiloxane scaffolds.The template protein rebound to the surface was measured more than twice as much ascompetitor. These scaffolds were then tested to understand their interaction with cells.The results of DNA and alkaline phosphatase activities indicate that the scaffolds thusdeveloped support growth and adhesion of osteoblastic cells. These initial selectivity andcytocompatibility studies show the potential of molecular-imprinted polysiloxanescaffolds to be used as tissue engineered materials for stable and controlled interactions atthe tissue-implant interface.