Thesis (Ph.D.)--Boston University / PLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at open-help@bu.edu. Thank you. / DNA sequencing techniques have evolved to the point where one can sequence millions of bases per minute, while our capacity to use this information has been left behind. One particularly notorious example is in the area of gene regulatory networks. A molecular study of gene regulation proceeds one protein at a time, requiring bench scientists months of work purifying transcription factors and performing DNA footprinting studies. Massive scale options like ChIP-Seq and microarrays are a step up, but still require considerable resources in terms of manpower and materials. While computational biologists have developed methods to predict protein function from sequence, gene locations from sequence, and even metabolic networks from sequence, the space of regulatory network reconstruction from sequence remains virtually untouched. Part of the reason comes from the fact that the components of a regulatory interaction, such as transcription factors and binding sites, are difficult to detect. The other, more prominent reason, is that there exists no "recognition code" to determine which transcription factors will bind which sites. I've created a pipeline to reconstruct regulatory networks starting from an unannotated complete genomic sequence for a prokaryotic organism. The pipeline predicts necessary information, such as gene locations and transcription factor sequences, using custom tools and third party software. The core step is to determine the likelihood of interaction between a TF and a binding site using a black box style recognition code developed by applying machine learning methods to databases of prokaryotic regulatory interactions. I show how one can use this pipeline to reconstruct the virtually unknown regulatory network of Bacillus anthracis. / 2999-01-01

Thesis (B.A.)--Boston University. University Professors Program Senior theses. / PLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at open-help@bu.edu. Thank you. / 2999-01-01

PEA3 is a member of the ETS family of genes. It is expressed in the adult mouse brain and epididymis. It is also expressed in the embryonal carcinoma P19 and F9 cell lines, where its expression is reduced following retinoic acid-induced differentiation. Sequencing of two human cDNA clones revealed the existence of alternatively-spliced PEA3 isoforms from the human PEA3 gene. Consequently there is a likelihood that such alternative splicing also occurs within the mouse PEA3 gene. If this is so, it would be of interest to determine the organization of the mouse PEA3 gene. This information will be necessary to explain any alternativelyspliced PEA3 isoforms that might be observed. To obtain this information, 21 kb of mouse genomic DNA containing the mouse PEA3 gene locus was isolated, subcloned and sequenced. This sequence data revealed that the mouse PEA3 messenger RNA (as present in several previously-characterized cDNAs) is encoded by 13 exons distributed over 14 kb of the mouse genome. To ensure that the complete PEA3 mRNA had been obtained and characterized, the 5' end of the PEA3 mRNA was isolated by the RACE protocol (Rapid Amplification of cDNA Ends). The results of this analysis indicated a transcription start site corresponding to the start of exon 1. The region immediately upstream of this likely initiation site contains numerous cognate sites for transcription factors which may play a role in controlling PEA3 transcription. / Thesis / Master of Science (MSc)

Thesis (Ph. D.)--University of Washington, 1999. / Vita. Includes bibliographical references (leaves 98-107).

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / A bactÃria Helicobacter pylori, à um agente etiolÃgico bem estabelecido para o desenvolvimento de lesÃes gÃstricas como gastrite, Ãlcera pÃptica, metaplasia e doenÃas malignas, com alta incidÃncia de infecÃÃo em todo mundo, porÃm cerca de 80% dos indivÃduos infectados permanecem assintomÃticos e apenas uma minoria desenvolve doenÃas a ela relacionadas. Estudos tÃm sido realizados na tentativa de identificar a relaÃÃo de genes determinantes da patogenicidade de H. pylori, entretanto, atà o momento apenas os genes cagA e o alelo de vacA s1m1 sÃo considerados marcadores de virulÃncia para o desenvolvimento de lesÃes gÃstricas mais graves. A bactÃria H. pylori possui uma alta variabilidade genÃtica, sendo que um dos mecanismos propostos para o desenvolvimento de lesÃo seria atravÃs da inflamaÃÃo. DiferenÃas na intensidade de respostas inflamatÃrias poderiam ser decorrentes do perfil genotÃpico da cepa infectante. Recentes trabalhos apontam a importÃncia dos genes cagE, virB11 de H. pylori, em cÃncer gÃstrico. Entretanto, estudos em lesÃes gÃstricas sÃo restritos. Assim, o objetivo deste trabalho foi determinar o perfil genotÃpico das cepas de H. pylori quanto a presenÃa dos genes de virulÃncia cagA, cagE, virB11, vacA e flaA, circulantes no estado do Cearà em 201 casos de lesÃes gÃstricas de diferentes gravidades, coletadas de pacientes dispÃpticos atendidos em trÃs hospitais de Fortaleza-CE. A detecÃÃo de H. pylori foi feita atravÃs da amplificaÃÃo do gene ureC, os genes estudados por amplificaÃÃo de fragmentos especÃficos, usando a tÃcnica de PCR, foram observados em gel de agarose 1% e poliacrilamida a 6% e 8%. Neste estudo houve um predomÃnio do sexo feminino 59% (119/201), principalmente na faixa etÃria (15-44) 30% (61/201), alta taxa de infecÃÃo por H. pylori 97,5% (196/201). A gastrite crÃnica ativa (GCA) foi a lesÃo gÃstrica mais frequente (55,7%;112/201), associada a pacientes na faixa etÃria de 15-44 (36,6% : 41/112) opondo-se à metaplasia intestinal em que, a frequÃncia dos casos aumentou com a idade sendo maior em pacientes mais velhos 46% (16/35), o que concorda com a literatura. O Ãnico caso de displasia ocorreu numa paciente > 65 anos. As lesÃes gÃstricas foram predominantemente localizadas no antro. O gene cagA foi mais frequente na gastrite crÃnica ativa (GCA), estando tambÃm em alta frequÃncia na gastrite atrÃfica (GA) e metaplasia intestinal (MI). Na GCA foi observado tambÃm alta frequÃncia dos genes cagE, virB11, vacAm1 e flaA, com diferenÃa estatÃstica, quando comparada com a gastrite crÃnica inativa (GCI) (cagA - p=0,007 cagE- p=0,000, virB11 - p=0,005, vacAm1- p=0,047 e flaA - p=0,001), sendo que os genes virB11 e flaA foram mais frequentes na GCA do que na Ãlcera. Os alelos s1 e m1 de vacA , bem como a combinaÃÃo s1m1 foram os mais frequentes nas lesÃes gÃstricas. Os casos H. pylori positivos foram agrupados de acordo com a presenÃa dos genes estudados, levando em consideraÃÃo a presenÃa do alelo s1 e os genes da ilha. Nessas anÃlises foi observada maior frequÃncia de cepas contendo os genes vacA s1 e [Ia (cagA+, cagE+ e virB11+ ) + Ib (cagA+ e virB11+; cagE+ e virB11+) ; 47,7% ] na gastrite crÃnica ativa em relaÃÃo a gastrite crÃnica inativa (GCI), esta com maior frequÃncia de cepas dos grupos [Ic (cagA+ ou cagE+ ou virB11+ ou cagA+ e cagE+) +Id (ureC+); 62%]. Adicionalmente, maior frequÃncia de cepas pertencentes aos grupos Ia e Ib foi observada na metaplasia intestinal incompleta, (59%), enquanto que na metaplasia intestinal completa uma maior frequencia de cepas pertencentes aos grupos Ic + Id, (78%) (p=0,027). Todos os casos de gastrite atrÃfica e Ãlcera eram do grupo I, e um Ãnico caso de displasia pertencia ao grupo Ia. Esses dados evidenciam uma associaÃÃo de cepas com genÃtipos mais virulentos em lesÃes com potencialidade de malignizaÃÃo. / A bactÃria Helicobacter pylori à um agente etiolÃgico bem estabelecido para o desenvolvimento de lesÃes gÃstricas como gastrite, Ãlcera pÃptica, metaplasia e doenÃas malignas, com alta incidÃncia de infecÃÃo todo mundo, porÃm cerca de 80% dos indivÃduos infectados permanecem assintomÃticos e apenas uma minoria desenvolve doenÃas a ela relacionadas. Estudos tem sido realizados na tentativa de identificar a relaÃÃo de genes determinantes da patogenicidade de H. pylori, entretanto, atà o momento apenas os genes cagA e o alelo de vacA s1m1 sÃo considerados marcadores de virulÃncia para o desenvolvimento de lesÃes gÃstricas mais graves. A bactÃria H. pylori possui uma alta variabilidade genÃtica sendo que um dos mecanismos propostos para o desenvolvimento de lesÃo seria atravÃs da inflamaÃÃo. DiferenÃas na intensidade de respostas inflamatÃrias poderiam ser decorrentes do perfil genotÃpico da cepa infectante. Recentes estudos apontam a importÃncia dos genes cagE, virB11 de H. pylori, em cÃncer gÃstrico. Entretanto, estudos em lesÃes gÃstricas sÃo restritos. Assim, o objetivo deste trabalho foi determinar o perfil genotÃpico das cepas de H. pylori quanto a presenÃa dos genes de virulÃncia cagA, cagE, virB11, vacA e flaA, circulantes no estado do Cearà em 201 casos de lesÃes gÃstricas de diferentes gravidades, coletadas de pacientes dispÃpticos atendidos em trÃs hospitais de Fortaleza-CE. A detecÃÃo de H. pylori foi feita atravÃs da amplificaÃÃo do gene ureC, e os genes estudados por amplificaÃÃo de fragmentos especÃficos, usando a tÃcnica de PCR, e foram observados em gel de agarose 1% e poliacrilamida a 6% e 8%. Nesse estudo houve um predomÃnio do sexo feminino 59% (119/201), principalmente na faixa etÃria (15-44) 30% (61/201), e alta taxa de infecÃÃo por H. pylori 97,5% (196/201). A gastrite crÃnica ativa (GCA), foi a lesÃo gÃstrica mais frequente (55,7%;112/201), associada a pacientes na faixa etÃria de 15-44 (36,6% : 41/112) opondo-se à metaplasia intestinal onde a frequÃncia dos casos aumentou com a idade sendo maior em pacientes mais velhos 46% (16/35), o que concorda com a literatura. O Ãnico caso de displasia ocorreu numa paciente > 65 anos. As lesÃes gÃstricas foram predominantemente localizadas no antro. O gene cagA foi mais frequente na gastrite crÃnica ativa (GCA), estando tambÃm em alta frequÃncia na gastrite atrÃfica (GA) e metaplasia intestinal (MI). Na GCA foi observado tambÃm alta freqÃÃncias dos genes cagE, virB11, vacAm1 e flaA com diferenÃa estatÃstica quando comparada com a gastrite crÃnica inativa (GCI) (cagA - p=0,007 cagE- p=0,000, virB11 - p=0,005, vacAm1- p=0,047 e flaA - p=0,001), sendo que os genes virB11 e flaA foram mais frequentes na GCA que na Ãlcera. Os alelos s1 e m1 de vacA , bem como a combinaÃÃo s1m1 foram os mais frequentes nas lesÃes gÃstricas. Os casos H. pylori positivos foram agrupados de acordo com a presenÃa dos genes estudados, levando em consideraÃÃo a presenÃa do alelo s1 e os genes da ilha. Nestas anÃlises foi observada maior frequÃncia de cepas contendo os genes vacA s1 e [Ia (cagA+, cagE+ e virB11+ ) + Ib (cagA+ e virB11+; cagE+ e virB11+) ; 47,7% ] na gastrite crÃnica ativa em relaÃÃo a gastrite crÃnica inativa (GCI), esta Ãltima com maior frequÃncia de cepas dos grupos [Ic (cagA+ ou cagE+ ou virB11+ ou cagA+ e cagE+) +Id (ureC+); 62%]. Adicionalmente, maior frequÃncia de cepas pertencentes aos grupos Ia e Ib foi observada na metaplasia intestinal incompleta, (59%), enquanto que na metaplasia intestinal completa uma maior frequencia de cepas pertencentes aos grupos Ic + Id, (78%) (p=0,027). Todos os casos de gastrite atrÃfica e Ãlcera eram do grupo I, e o Ãnico caso de displasia pertencia ao grupo Ia. Esses dados evidenciam uma associaÃÃo de cepas com genÃtipos mais virulentos em lesÃes com potencialidade de malignizaÃÃo. / The bacterium Helicobacter pylori is a well-established etiological factor in the development of gastric lesions such as gastritis, peptic ulcers, metaplasia and malignancy. The incidence of infection by this pathogen is high worldwide, but about 80% of infected individuals remain asymptomatic, and only a minority develops related diseases. Many studies have been conducted in an attempt to identify the involvement of genes in determining the pathogenicity of H. pylori, but so far, only the genes cagA and vacA allele s1m1 are considered virulence markers for the development of the more severe gastric lesions. H. pylori bacteria have a high genetic variability and one of the proposed mechanisms for lesion development is through inflammation. Differences in the intensity of inflammatory responses could be due to the genotypic profile of the infecting strain. Recent studies indicate the importance of the genes cagE and virB11, but mostly involving gastric cancer, while their role in gastric lesions is limited. The objective of this study was to determine the genetic subtypes of H. pylori strains and the presence of the virulence genes cagA, cagE, virB11 and flaA genes and vacA alleles, circulating in Ceara state, Brazil. Samples were collected from 201 cases of gastric lesions of varying severity, in dyspeptic patients treated at three hospitals in Fortaleza, Ceara State. The detection of H. pylori was performed using ureC gene amplification by PCR, and the detection of the genes studied was carried out by amplification of gene-specific fragments separated in 1% agarose gels. The sample was predominantly female (59%, 119/201) and mainly in the age group 15-44 years old (30%, 61/201), and had a high rate of H. pylori infection (97.5%, 196/201). Active chronic gastritis (ACG) was the most common gastric lesion (55.7%, 112/201), associated with patients aged 15-44 (36.6%, 41/112), unlike intestinal metaplasia, in which the frequency of cases increased with age, being higher in older patients (46%, 16/35), which agrees with the literature. A single case of dysplasia occurred in a patient > 65 years. The lesions were predominantly located in the gastric antrum and 30% of the cases had lesions located in the body and antrum simultaneously. The cagA gene was more frequent in ACG, showing a statistically significant correlation (r = 0.220, p = 0.007); it also showed a high frequency in atrophic gastritis and in intestinal metaplasia. In ACG, a high frequency of the genes cagE, virB11, flaA and vacAm1 was also statistically associated when compared to chronic inactive gastritis (ICG) (cagA - p = 0.007, cagE - p = 0.000, virB11 - p = 0.005, vacAm1 p = 0.047 and flaA - p = 0.001), and the genes virB11 and flaA were more frequent in ACG than in ulcer. The vacA alleles s1 and m1, and the combination s1m1, were the most frequent ones in gastric lesions. Considering the genotypes of H. pylori grouped by the presence of the genes studied, we observed a higher frequency of the most virulent strains in the ACG group (Ia + Ib, 47.7%) when compared to ICG, the latter showing a higher frequency of less virulent strains (Ic + Id, 62%). Additionally, a higher frequency of more virulent strains, belonging to groups Ia and Ib, was observed in incomplete intestinal metaplasia (59%), while in complete intestinal metaplasia an increased frequency of less virulent strains, belonging to the groups Ic + Id (78%) (p = 0.027), was found. All cases of atrophic gastritis and ulcer were in group I, and the single case of dysplasia belonged to group Ia (high virulence). These data indicate the important role of more virulent strains in potential malignant lesions.

Background Elevated concentrations of low-density lipoprotein cholesterol (LDL-c) are a major risk factor for the development of coronary artery disease (CAD) because of their role in the progression of atherosclerosis. The black South African population is known to have had historically low LDL-c and in the past there was almost no CAD in the population. However, as this population moves through the nutrition transition, LDL-c levels are increasing. LDL-c levels are regulated by the LDL receptors, which is the major protein involved with transporting cholesterol across cell membranes in humans. Proprotein convertase subtilisinlike/kexin type 9 (PCSK9) is another protein involved with the regulation of LDL-c through its role in assisting with the degradation of the LDL receptor. Variants in both genes can cause elevated or lowered LDL-c levels. Very little information is available on the frequency or presence of variants in the low-density lipoprotein receptor (LDLR) and PCSK9 gene in the black South African population and on how these variants associate with LDL-c. The main aim of the study was thus to determine novel and existing genetic variants in these two genes and to describe the manner in which they associate with plasma LDL-c levels in a black South African population undergoing an epidemiological transition. Methods The 2005 baseline data from the Prospective Urban and Rural (PURE) study population were used in this study. The study population consisted of apparently healthy black volunteers form the North West province of South Africa, aged 35 to 60 years. Thirty individuals were randomly chosen from the 1860 volunteers to determine the presence of known and novel variants in these genes by automated bidirectional sequencing. The promoter region, exons and flanking regions were sequenced and variants were identified utilising CLC DNA Workbench. Deoxyribonucleic acid (DNA) samples for 1500 individuals of the PURE study population were genotyped by means of a Golden Gate Genotyping Assay. Analyses of covariance (ANCOVA) were used to test for associations between the different genotypes in both the LDLR and PCSK9 genes and LDL-c levels. Haplotypes were generated by using the confidence intervals on the software programme, HaploView. A genetic risk score (GRS) was determined by including variants which associated significantly with LDL-c. The GRS, the haplotypes and the variants that associated significantly with LDL-c were used in separate linear regression models with variants which correlated with LDL-c to determine how all these variables contribute to the differences in LDL-c levels. Results and discussion Novel and known variants were identified in both the genes and in total 52 variants were genotyped. Rare variants such as rs17249141 and rs28362286 were detected in the study population and are associated with low levels of LDL-c. The variants identified in the LDLR gene were situated largely in regulatory regions such as the promoter, intron and 3‟untranslated regions. Haplotypes in the LDLR gene with the highest frequency associated with lower LDL-c levels, which could contribute to the study population‟s low mean LDL-c level. Haplotypes identified in the PCSK9 gene had a weaker association with LDL-c levels. The minor allele frequencies of many of the variants differed from those of the European population and therefore the importance of population-specific research cannot be sufficiently emphasised. The GRS, haplotypes and variants used in the regression models to determine whether they contributed to predicting the variance in LDL-c in the study population made a small contribution to explaining this. BMI best explained the variance in LDL-c levels. Older women with a body mass index (BMI)>25kg/m2 were identified as being at greater risk of developing elevated LDL-c levels than the rest of the study population. Heterozygote carriers of variant, rs28362286, had 0.787 mmol/L lower LDL-c than carriers of the wild type and this is associated with a reduced risk of developing CAD. Conclusion and recommendation When considering the results mentioned above, adding genetic analysis to explaining the variance in LDL-c levels seems to have its limitations, but the study included only two of many genes that play a role in the metabolism and regulation of LDL-c levels. Incorporating more genes and more variants into analyses and prediction models will add greater value to defining LDL-c levels. Rarer variants with a large impact on protein function, such as rs28362286, have a greater effect on LDL-c levels and could predict the variance better than the common variants. Risk factors such as BMI can also still be trusted to indicate which individuals or groups are at risk of developing elevated LDL-c levels. Health advice should be given to appropriate target groups such as older women with a BMI >25kg/m2 in order to prevent CAD from becoming a burden in this population. / PhD (Dietetics), North-West University, Potchefstroom Campus, 2014

Background Elevated concentrations of low-density lipoprotein cholesterol (LDL-c) are a major risk factor for the development of coronary artery disease (CAD) because of their role in the progression of atherosclerosis. The black South African population is known to have had historically low LDL-c and in the past there was almost no CAD in the population. However, as this population moves through the nutrition transition, LDL-c levels are increasing. LDL-c levels are regulated by the LDL receptors, which is the major protein involved with transporting cholesterol across cell membranes in humans. Proprotein convertase subtilisinlike/kexin type 9 (PCSK9) is another protein involved with the regulation of LDL-c through its role in assisting with the degradation of the LDL receptor. Variants in both genes can cause elevated or lowered LDL-c levels. Very little information is available on the frequency or presence of variants in the low-density lipoprotein receptor (LDLR) and PCSK9 gene in the black South African population and on how these variants associate with LDL-c. The main aim of the study was thus to determine novel and existing genetic variants in these two genes and to describe the manner in which they associate with plasma LDL-c levels in a black South African population undergoing an epidemiological transition. Methods The 2005 baseline data from the Prospective Urban and Rural (PURE) study population were used in this study. The study population consisted of apparently healthy black volunteers form the North West province of South Africa, aged 35 to 60 years. Thirty individuals were randomly chosen from the 1860 volunteers to determine the presence of known and novel variants in these genes by automated bidirectional sequencing. The promoter region, exons and flanking regions were sequenced and variants were identified utilising CLC DNA Workbench. Deoxyribonucleic acid (DNA) samples for 1500 individuals of the PURE study population were genotyped by means of a Golden Gate Genotyping Assay. Analyses of covariance (ANCOVA) were used to test for associations between the different genotypes in both the LDLR and PCSK9 genes and LDL-c levels. Haplotypes were generated by using the confidence intervals on the software programme, HaploView. A genetic risk score (GRS) was determined by including variants which associated significantly with LDL-c. The GRS, the haplotypes and the variants that associated significantly with LDL-c were used in separate linear regression models with variants which correlated with LDL-c to determine how all these variables contribute to the differences in LDL-c levels. Results and discussion Novel and known variants were identified in both the genes and in total 52 variants were genotyped. Rare variants such as rs17249141 and rs28362286 were detected in the study population and are associated with low levels of LDL-c. The variants identified in the LDLR gene were situated largely in regulatory regions such as the promoter, intron and 3‟untranslated regions. Haplotypes in the LDLR gene with the highest frequency associated with lower LDL-c levels, which could contribute to the study population‟s low mean LDL-c level. Haplotypes identified in the PCSK9 gene had a weaker association with LDL-c levels. The minor allele frequencies of many of the variants differed from those of the European population and therefore the importance of population-specific research cannot be sufficiently emphasised. The GRS, haplotypes and variants used in the regression models to determine whether they contributed to predicting the variance in LDL-c in the study population made a small contribution to explaining this. BMI best explained the variance in LDL-c levels. Older women with a body mass index (BMI)>25kg/m2 were identified as being at greater risk of developing elevated LDL-c levels than the rest of the study population. Heterozygote carriers of variant, rs28362286, had 0.787 mmol/L lower LDL-c than carriers of the wild type and this is associated with a reduced risk of developing CAD. Conclusion and recommendation When considering the results mentioned above, adding genetic analysis to explaining the variance in LDL-c levels seems to have its limitations, but the study included only two of many genes that play a role in the metabolism and regulation of LDL-c levels. Incorporating more genes and more variants into analyses and prediction models will add greater value to defining LDL-c levels. Rarer variants with a large impact on protein function, such as rs28362286, have a greater effect on LDL-c levels and could predict the variance better than the common variants. Risk factors such as BMI can also still be trusted to indicate which individuals or groups are at risk of developing elevated LDL-c levels. Health advice should be given to appropriate target groups such as older women with a BMI >25kg/m2 in order to prevent CAD from becoming a burden in this population. / PhD (Dietetics), North-West University, Potchefstroom Campus, 2014

Cell shape and motility are primarily controlled by cellular mechanics. The attachment of the plasma membrane to the underlying actomyosin cortex has been proposed to be important for cellular processes involving membrane deformation. However, little is known about the actual function of membrane-to-cortex attachment (MCA) in cell protrusion formation and migration, in particular in the context of the developing embryo. Here, we use a multidisciplinary approach to study MCA in zebrafish mesoderm and endoderm (mesendoderm) germ layer progenitor cells, which migrate using a combination of different protrusion types, namely, lamellipodia, filopodia, and blebs, during zebrafish gastrulation. By interfering with the activity of molecules linking the cortex to the membrane and measuring resulting changes in MCA by atomic force microscopy, we show that reducing MCA in mesendoderm progenitors increases the proportion of cellular blebs and reduces the directionality of cell migration. We propose that MCA is a key parameter controlling the relative proportions of different cell protrusion types in mesendoderm progenitors, and thus is key in controlling directed migration during gastrulation.

Clonally derived liver tumours and an ovarian cyst developed in mice following EIAV and FIV delivery in utero. LAM PCR and 454 sequencing was used to retrieve proviral insertion sites. TaqMan analysis revealed gene expression changes in lentiviral infected tumours. STRING and IPA networks identified links between genes flanking the lentivirus provirus and oncogenic pathways supporting the role of insertional mutagenesis in Hepatocellular Carcinoma (HCC). Global methylation analysis demonstrated increased relative methylation levels in lentivirus (EIAV, FIV, and HIV) infected normal and tumour samples. This provided strong evidence for host defence against lentivirus infection by epigenetic means. Microarray data showed altered expression of Dnmt1 and Dnmt3b and TaqMan analysis revealed specific changes in Dnmts levels when compared to uninfected liver. The evidence found for involvement of DNA methylation associated with lentivirus infection and possibly tumour development required that this study be repeated in vitro. DNA methylation was investigated at early time points after lentivirus and retrovirus infection in HepG2 cells. Results revealed sharp increases in global methylation and Dnmt levels at 24 and 30hrs post infection. E2F targets play a key role in the regulation of gene expression and aberrations result in the development of cancer. Of the 94 E2F target genes analysed 77.7% were involved in DNA damage and repair mechanisms, 21.3% were known oncogenes or shown to exert oncogenic activity and 80.9% were categorised as HCC target genes. The fact that all lentiviral/retroviral vectors used in this study were found to cause changes in methylation and gene expression in vivo and in vitro suggests that these vectors, at least in the mouse, are genotoxic. Findings here support the use of the fetal animal model to identify vector genotoxicity and the mechanisms of lentiviral vector-induced tumorigenesis. This model may be a valuable tool to evaluate the safety of lentiviruses for gene therapy.

Using a generalized-clique hidden Markov model (HMM) as the starting point for a eukaryotic gene finder, the objective here is to strengthen the signal information at the transitions between coding and non-coding (c/nc) regions. This is done by enlarging the primitive hidden states associated with individual base labeling (as exon, intron, or junk) to substrings of primitive hidden states or footprint states. Moreover, the allowed footprint transitions are restricted to those that include either one c/nc transition or none at all. (This effectively imposes a minimum length on exons and the other regions.) These footprint states allow the c/nc transitions to be seen sooner and have their contributions to the gene-structure identification weighted more heavily – yet contributing as such with a natural weighting determined by the HMM model itself according to the training data – rather than via introducing an artificial gain-parameter tuning on major transitions. The selection of the generalized HMM model is interpolated to highest Markov order on emission probabilities, and to highest Markov order (subsequence length) on the footprint states. The former is accomplished via simple count cutoff rules, the latter via an identification of anomalous base statistics near the major transitions using Shannon entropy. Preliminary indications, from applications to the C. elegans genome, are that the sensitivity/specificity (SN/SP) result for both the individual state and full exon predictions are greatly enhanced using the generalized-clique HMM when compared to the standard HMM. Here the standard HMM is represented by the choice of the smallest size of footprint state in the generalized-clique HMM. Even with these improvements, we observe that both extremely long and short exon and intron segments would go undetected without an explicit model of the duration of state. The key contributions of this effort are the full derivation and experimental confirmation of a rudimentary, yet powerful and competitive gene finding method based on a higher order hidden Markov model. With suitable extensions, this method is expected to provide superior gene finding capability – not only in the context of pre-conditioned data sets as in the evaluations cited but also in the wider context of less preconditioned and/or raw genomic data.