<p>To investigate mechanisms involved in virus-host interactions, global changes in host gene expression were examined during infection with adenovirus type 2 (Ad2) using cDNA microarray technology. </p><p>In paper I and II, transcriptional changes in HeLa cells were investigated during the early and late phase of infection, respectively. A limited number of genes, mainly implicated in cell growth and antiviral defence, were found to be differentially expressed in the early phase, whereas modulation of host cell gene expression during the late phase was augmented and mainly focused on growth inhibition and cell architecture. </p><p>The experimental set-up was then redesigned to follow transcriptional regulatory events in growth synchronised, human primary lung fibroblasts. The immediate response of the host cell within two hours of infection was investigated in paper III, revealing a transient induction of a small number of cellular alert genes. This was followed by an expanded time course presented in paper IV, which included gene expression profiling at eight consecutive time points throughout the infectious cycle. The results indicated that specific sets of cellular genes were targeted at different stages of the infection, and four distinct periods were identified. </p><p>In summary, the studies presented in this thesis demonstrate that adenovirus interferes with many cellular processes during the progression of infection to optimize the cellular environment for viral replication. These include cell cycle control, cell growth and growth inhibition, as well as DNA, RNA and protein metabolism. However, a transient induction of cellular genes involved in immune response and growth inhibition was observed before the onset of viral gene expression. During the very late stages of infection, the expression of a large number of genes involved in maintaining the cell structure was down-regulated, presumably to facilitate the spread of progeny virus.</p>

<p>Osteoporosis is a disease that leads to decreased bone mineral density (BMD), an altered bone micro-architecture and fragile bones. The disease is highly heritable and numerous genes are thought to be involved, making it difficult to identify the causative genetic elements.</p><p>Animal models, mainly intercrosses between laboratory strains of mice, have been succesfully used to map genes affecting these traits, but may not mirror the multifactorial genetic etiology of highly complex traits such as osteoporosis.</p><p>Over the course of tens of thousand years humans have kept domestic animals whose phenotypic repertoires have been tailored to meet our needs. Wild-type red junglefowl (RJ) and domestic White Leghorn (WL) chicken differ for several bone traits. </p><p>In this thesis Quantitative Trait Loci (QTL) mapping was used to trace the inheritance of bone traits in two separate intercrosses between RJ and WL. In these studies we identified several QTL that contributed to differences in BMD, bone size and biomechanical strength of bone. In a comparison of QTL identified in the two intercrosses it was observed that nine QTL had overlapping genomic positions, implicating these loci as important to bone phenotypic variation in chicken.</p><p>In two separate studies, microarray technology was used to compare global gene expression in bone tissue from RJ and WL. In these studies, differential expression was observed for 779 and 560 genes, respectively. Many differentially expressed genes were co-localized with QTL, which implicates them as QTL-candidates. </p><p>Results presented in this thesis link several genomic regions and genes to variation in bone traits. Increased knowledge about these identified genes and regions will contribute to a better understanding of the mechanisms underlying inter-individual differences in bone metabolism, both in chicken and man.</p>

Fueled by ever-growing genomic information and rapid developments of proteomics–the large scale analysis of proteins and mapping its functional role has become one of the most important disciplines for characterizing complex cell function. For building functional linkages between the biomolecules, and for providing insight into the mechanisms of biological processes, last decade witnessed the exploration of combinatorial and chip technology for the detection of bimolecules in a high throughput and spatially addressable fashion. Among the various techniques developed, the protein chip technology has been rapid. Recently we demonstrated a new platform called “Spacially addressable protein array” (SAPA) to profile the ligand receptor interactions. To optimize the platform, the present study investigated various parameters such as the surface chemistry and role of additives for achieving high density and high-throughput detection with minimal nonspecific protein adsorption. In summary the present poster will address some of the critical challenges in protein micro array technology and the process of fine tuning to achieve the optimum system for solving real biological problems. / Singapore-MIT Alliance (SMA)

The occurrence of blooms of toxic cyanobacteria in freshwater environments is a global ecological and public health concern. Species of Microcystis are of particular importance because blooms occur in many freshwater environments throughout the world and microcystin toxin concentrations can exceed World Health Organization advisory levels. While microcystin has been associated with fish kills, sublethal effects of chronic exposure at environmentally relevant concentrations are relatively unknown. The objective of this research was to evaluate toxicity of microcystin and Microcystis in fish during all life history stages. We evaluated global gene expression response in larval zebrafish (Danio rerio), and a sub-set of biomarker genes indicative of microcystin exposure were identified. In addition, vitellogenin genes were highly up-regulated in zebrafish exposed to Microcystis but not the microcystin toxin, indicating potential endocrine disrupting effects of Microcystis blooms. Effects on reproduction were evaluated in adult zebrafish exposed to Microcystis. There was a significant decrease in the percentage of adults that spawned, however fecundity and larval survival were not affected. Laboratory mesocosm experiments with channel catfish (Ictalurus punctatus) were also conducted to determine the importance of dietary and aqueous exposure in microcystin bioaccumulation and assess histopathological lesions. Tissue toxin concentrations and histopathological lesions were also evaluated in channel catfish collected from Lake Erie and Waterville Reservoir, North Carolina to monitor fish living in environments affected by Microcystis blooms and relate responses to those observed in laboratory exposures.

Statisticians are used to model interactions between variables on the basis of observed data. In a lot of emerging fields, like bioinformatics, they are confronted with datasets having thousands of variables, a lot of noise, non-linear dependencies and, only, tens of samples. The detection of functional relationships, when such uncertainty is contained in data, constitutes a major challenge. Our work focuses on variable selection and network inference from datasets having many variables and few samples (high variable-to-sample ratio), such as microarray data. Variable selection is the topic of machine learning whose objective is to select, among a set of input variables, those that lead to the best predictive model. The application of variable selection methods to gene expression data allows, for example, to improve cancer diagnosis and prognosis by identifying a new molecular signature of the disease. Network inference consists in representing the dependencies between the variables of a dataset by a graph. Hence, when applied to microarray data, network inference can reverse-engineer the transcriptional regulatory network of cell in view of discovering new drug targets to cure diseases. In this work, two original tools are proposed MASSIVE (Matrix of Average Sub-Subset Information for Variable Elimination) a new method of feature selection and MRNET (Minimum Redundancy NETwork), a new algorithm of network inference. Both tools rely on the computation of mutual information, an information-theoretic measure of dependency. More precisely, MASSIVE and MRNET use approximations of the mutual information between a subset of variables and a target variable based on combinations of mutual informations between sub-subsets of variables and the target. The used approximations allow to estimate a series of low variate densities instead of one large multivariate density. Low variate densities are well-suited for dealing with high variable-to-sample ratio datasets, since they are rather cheap in terms of computational cost and they do not require a large amount of samples in order to be estimated accurately. Numerous experimental results show the competitiveness of these new approaches. Finally, our thesis has led to a freely available source code of MASSIVE and an open-source R and Bioconductor package of network inference.

Head and neck cancer is the 6th most common malignancy worldwide, with tumours of the tongue being one of the most prevalent sites. Despite advances in surgery and radiotherapy, the five-year survival has not changed during the last decades and remains at approximately 50%. Identification of novel biomarkers for more personalized treatment is important for increasing survival in these patients. One of the most commonly used methods in the search for new biomarkers is microarray analysis. A substantial limitation with this technique is the requirement for fresh frozen samples from which high quality RNA can be extracted. This becomes particularly problematic when attempting to discover differences associated with individual sub-types or rare cancers. Recent developments, including the DASL microarray platform, have provided the possibility of analysing RNA of poorer quality from formalin fixed paraffin embedded (FFPE) samples. FFPE is the standard way of preserving tissue from patients and millions of samples are stored around the world. In this thesis we have evaluated the use of FFPE samples and global gene expression profiling for increasing basic knowledge in a subgroup of oral cancer patients with tumours of the tongue. As confirmation of microarray results using qPCR is of outmost importance for conclusive data evaluation, we first aimed at finding a housekeeping gene stably expressed across malignant and non-malignant FFPE oral tissue. TUBA6, which belongs to the tubulin family was detected as being the most stable out of eight possible genes and was thus used for qPCR normalization throughout the following studies. We have performed three separate microarray experiments. Initially only a focused DASL array covering 502 cancer related genes was available and we used it to analyze a smaller cohort of patients and controls (n=36). A similar cohort (n=29) was also analyzed for expression of 836 micoRNAs. In 2009 a whole genome DASL array was launched, covering over 20,000 genes, and all tongue tumour samples available between 1997 and 2010 (n=87) were analysed using this array. Similar to other research groups we observed very high replicate reproducibility using both DASL arrays. When using the microRNA array and the whole genome DASL array an effect of sample quality on the detected expression level of individual genes was noticed. While the expression of some genes severely decreased with a decrease in sample quality others were not changed. This will impair normalization, leading to a residual non-biological variation within the data. Based on our findings we have presented some recommendations for minimizing the effect of sample quality and maximizing the level of biologically relevant information obtained from these experiments, e.g. ensuring that samples in groups to be compared are of the same quality range. For the microRNA data we also introduced an additional normalization step to the standard normalizations. We could show that lists of differentially expressed genes generated when taking these precautions were enriched for genes involved in cancer related processes and contained for tongue carcinoma previously identified changes. A number of differentially expressed genes, novel for tongue carcinoma, were also confirmed in high quality fresh frozen samples, including BCL2A1 (apoptosis), CXCL10 (immune response), SLC2A6 (energy transport) and miR-424 (angiogenesis). In conclusion microarrays can be used to analyze FFPE samples but should be performed with care. Standard normalization methods will not remove the variation introduced by samples being of different quality, leading to spurious results. Taking a few precautions, however, led to the identification of differentially expressed genes relevant in tumour development and maintenance. The recommendations we make can facilitate design of future studies using FFPE samples. The genes we identified as being differentially expressed in tumour tissue now need to be further evaluated for their potential as biomarkers in tongue carcinoma.

<p>Autism is a pervasive developmental disorder (PDD) that&rsquo / s incidence is approximately 1 in 158. It is four times more prevalent in males than females and is believed to be caused by both genetic and environmental factors. Research indicates that several genes are involved in autism and it is believed that these genes act together to produce autism. Many genes implicated in this disorder are involved with brain structure formation and brain functioning. Studies have identified the reelin (RELN) gene as necessary for proper formation of brain, which indicates that RELN abnormalities could contribute to the aetiology of several neurogenetic diseases such as schizophrenia, bipolar and autism. The aims of the study were (i) to genotype two SNPs (exonic rs3622691 and intronic rs736707) in the RELN gene using Taqman&reg / SNP Genotyping assays to detect association with autism in three distinct South African (SA) ethnic groups (Black, Caucasian and Mixed), and (ii) to detect candidate genes that are over and under-expressed in the samples taken from a SA Caucasian autistic group and compare those with samples taken from a healthy Caucasian group using cDNA microarray. The Taqman&reg / study indicated significant association for the intronic SNP, rs736707, with a p-value of 0.0009 in the total SA group. More so, the Mixed group displayed the highest significance amongst the ethnic groups, with a p-value of 0.00014. The microarray study yielded 21 genes with 95% significance in the Caucasian sample group. Most genes were hypothetical proteins and formed part of the FAM90A family. The LOC83459 showed the highest level of expression in the autistic samples, while the BTNL8 gene was shown to be highly suppressed in the control samples.</p>

In this thesis we attempted to elicit patterns of gene expression that influence phenotype, and that may also have been altered by thousands of years of domestication and selection, between red junglefowl and White Leghorn chickens. Red junglefowl are the wild ancestor to all domesticated chickens, and poultry in general are highly valued as a research animal and food resource. The project was also begun in order to complement an earlier study of an intercross between White Leghorn and red junglefowl, which identified several regions that were linked with phenotypic differences between the two birds. We began by creating our own cDNA microarray via generating four cDNA libraries from red junglefowl/White Leghorn brain and testis. We generated 12,549 unique transcripts. This included 400 new putative transcripts specific to chickens, and 180 transcripts that were not found in any other database. When investigating polymorphisms between White Leghorn and red junglefowl we found a SNP rate of 1.9/kb coding region, and a synonymous and non-synonymous percentage for these SNPs of 80 and 20% respectively. In the last two studies we used the cDNA microarray to measure gene expression differences between White Leghorn and red junglefowl in both hypothalamus/thalamus and liver. We found that there appears to be a significant number of genes down-regulated in White Leghorn hypothalamus/thalamus, plus an over-representation of up-regulated genes from well-known pathways, as compared with red junglefowl. We hypothesize that domestication/selection may be connected with this characteristic. We also found that the p-arm of chicken chromosome 4, which is an ancestral microchromosome, was over represented with differentially expressed genes in hypothalamus/thalamus. A number of differentially expressed genes are shared between the two tissues, and these genes are expressed in same manner between red junglefowl and White Leghorn.

To investigate mechanisms involved in virus-host interactions, global changes in host gene expression were examined during infection with adenovirus type 2 (Ad2) using cDNA microarray technology. In paper I and II, transcriptional changes in HeLa cells were investigated during the early and late phase of infection, respectively. A limited number of genes, mainly implicated in cell growth and antiviral defence, were found to be differentially expressed in the early phase, whereas modulation of host cell gene expression during the late phase was augmented and mainly focused on growth inhibition and cell architecture. The experimental set-up was then redesigned to follow transcriptional regulatory events in growth synchronised, human primary lung fibroblasts. The immediate response of the host cell within two hours of infection was investigated in paper III, revealing a transient induction of a small number of cellular alert genes. This was followed by an expanded time course presented in paper IV, which included gene expression profiling at eight consecutive time points throughout the infectious cycle. The results indicated that specific sets of cellular genes were targeted at different stages of the infection, and four distinct periods were identified. In summary, the studies presented in this thesis demonstrate that adenovirus interferes with many cellular processes during the progression of infection to optimize the cellular environment for viral replication. These include cell cycle control, cell growth and growth inhibition, as well as DNA, RNA and protein metabolism. However, a transient induction of cellular genes involved in immune response and growth inhibition was observed before the onset of viral gene expression. During the very late stages of infection, the expression of a large number of genes involved in maintaining the cell structure was down-regulated, presumably to facilitate the spread of progeny virus.

Osteoporosis is a disease that leads to decreased bone mineral density (BMD), an altered bone micro-architecture and fragile bones. The disease is highly heritable and numerous genes are thought to be involved, making it difficult to identify the causative genetic elements. Animal models, mainly intercrosses between laboratory strains of mice, have been succesfully used to map genes affecting these traits, but may not mirror the multifactorial genetic etiology of highly complex traits such as osteoporosis. Over the course of tens of thousand years humans have kept domestic animals whose phenotypic repertoires have been tailored to meet our needs. Wild-type red junglefowl (RJ) and domestic White Leghorn (WL) chicken differ for several bone traits. In this thesis Quantitative Trait Loci (QTL) mapping was used to trace the inheritance of bone traits in two separate intercrosses between RJ and WL. In these studies we identified several QTL that contributed to differences in BMD, bone size and biomechanical strength of bone. In a comparison of QTL identified in the two intercrosses it was observed that nine QTL had overlapping genomic positions, implicating these loci as important to bone phenotypic variation in chicken. In two separate studies, microarray technology was used to compare global gene expression in bone tissue from RJ and WL. In these studies, differential expression was observed for 779 and 560 genes, respectively. Many differentially expressed genes were co-localized with QTL, which implicates them as QTL-candidates. Results presented in this thesis link several genomic regions and genes to variation in bone traits. Increased knowledge about these identified genes and regions will contribute to a better understanding of the mechanisms underlying inter-individual differences in bone metabolism, both in chicken and man.