Human embryonic stem cells are derived from the pre-implantation blastocyst, residing in an extremely low oxygen environment. Application of this in vivo oxygen concentration to an in vitro setting has previously shown to be essential for driving the self-renewal and directed differentiation of human pluripotent stem cells. However, studies on hESCs at low oxygen have been frequently contra-indicatory. Here, comparative analysis of a hESC and iPSC line at a low oxygen concentration resulted in divergent effects across the two cell lines. Augmented TGFβ signalling was observed in conjunction with up-regulated transcription of pluripotency markers NANOG and POU5F1 in hESCs but not iPSCs. IPSCs also shifted to a state of increased proliferation whereas hESCs did not. It was also observed that exposure of hESCs and iPSCs to light throughout low oxygen culture induced large amounts of apoptosis, highlighting the requirement for careful selection of cell culture equipment for environmental oxygen control. Both the embryonic and adult brain retain tissue specific oxygen concentrations far below that at atmospheric oxygen concentrations. Previous studies showed that embryonic silencing of factors responsive to low oxygen caused a range of embryonic abnormalities, including deformation of the neural plate and tube. Here, differentiation of hESCs to an early neuroectodermal identity at low O2 did not definitively augment production of NPCs but did additively suppress BMP signalling above that at atmospheric O2. A concordant rise in apoptosis was also observed in a manner both independent of, and augmented by, exogenous BMP inhibition. Subsequently, neurogenesis at low oxygen produced terminal neurons with accelerated and augmented synaptic and induced excitability. This was characterised by a rise in the rate of membrane depolarisation, increased action potential overshoot and accelerated expression of pre-synaptic marker Synaptophysin. These results highlight a novel, critical role for low oxygen in augmenting the excitability of hESC derived neurons.

The DEFA1A3 locus on human chromosome 8p23.1 exhibits extensive copy number variation; individuals have between 3-16 copies of DEFA1A3. The region has additional complexity in that each repeat unit contains a gene locus that can be occupied by one of two different genes, DEFA1 or DEFA3. These encode the human neutrophil peptides (HNPs) 1-3, antimicrobial peptides involved in the innate immune response. In order to understand the mutational processes and evolutionary history of a complex locus like DEFA1A3, spatial information is essential. Whilst haplotype DEFA1A3 copy numbers and haplotype ratios of DEFA1 vs. DEFA3 have been determined, little is known about the features shared by, and the structures of, related haplotypes. In this study, flanking sequence variation has been used to identify five classes of DEFA1A3 haplotype, which are tagged by four SNPs. Haplotypes within each class share similar features, such as DEFA1A3 copy number, but the associations differ between-class and between-population. Emulsion haplotype fusion-PCR has been used to determine the spatial arrangement of the DEFA1 and DEFA3 genes, as well as additional internal variants, across haplotypes of European ancestry. A comparison of the structures of related haplotypes suggests that the predominant mechanism of copy number change at the DEFA1A3 locus is intra-allelic rearrangements (i.e. between haplotypes from the same class), facilitated by the high sequence similarity of repeat units within each class. This explains the preservation of linkage disequilibrium across the DEFA1A3 locus. The relationship between DEFA1A3 copy number and gene expression is unclear. A comparison between DEFA1A3 haplotype class and HNP1-3 expression in a UK cohort suggests that DEFA1A3 haplotype structure does not influence gene expression. However, the identification of four SNPs which tag DEFA1A3 haplotype class and, in turn, haplotype structure in haplotypes of European ancestry, will aid further studies in this area.

In the past decade, limited progress has been made in identifying genetic associations with performance and health-related fitness phenotypes due to the use primarily of the traditional candidate-gene approach involving small sample sizes and few coordinated research efforts. Much of the genetic data relating to human performance has been generated while exploring the aetiology of lifestyle-related disorders such as obesity and type 2 diabetes mellitus (T2DM). As of 2008, over 200 autosomal gene entries and quantitative trait loci have been reported to be significantly associated with performance and health-related fitness. However, most genetic findings to date have been inconclusive due to studies employing relatively small sample sizes and predominantly single-gene approaches which are especially prone to type I errors. It is widely accepted that there will be many genes involved in sporting performance and health-related fitness phenotypes, and hence it is timely that genetic research has moved to the genomics era with the use of a genome-wide approach (e.g. genotyping a large number of variants simultaneously across the entire human genome) in a well-phenotyped, large cohort. This thesis summarizes the recent findings of genetic predisposition to elite human performance by using the conventional candidate-gene approach as well as the unbiased genome-wide approach (i.e. genome-wide association studies, GWASs). The current candidate gene study focused on investigating whether polymorphisms in the angiotensin-converting enzyme (ACE) and α-actinin-3 (ACTN3) genes are associated with elite swimmer status (stratified by swimming distance) in Caucasians and East Asians. ACE I/D and ACTN3 p.R577X polymorphisms were genotyped for 200 elite Caucasian swimmers (short and middle distance, ≤ 400 m, n = 130; long distance, > 400 m, n = 70) and 326 elite Japanese and Taiwanese swimmers (short distance, ≤ 100 m, n = 166; middle distance, 200–400 m, n = 160). Logistic regression and multiple-testing adjustment were applied to test for these genetic associations. ACE I/D was found to be associated with swimmer status in Caucasians, with the D allele being overrepresented in short-and-middle-distance swimmers with the largest effect being observed for the I-allele-dominant model (odds ratio = 1.90; logistic regression p = 0.001; permutation test p = 0.0005). In East Asians, however, the I allele was overrepresented in the short-distance swimmer group under the D-allele-dominant model (odds ratio = 1.52; logistic regression p = 0.012; permutation test p = 0.0098). The ACE I/D association findings in the elite swimmer cohorts showed that different risk alleles responsible for the associations were observed in swimmers of different ethnicities. ACTN3 p.R577X was not statistically significantly associated with swimmer status in either Caucasian or East Asian population. The lack of associations between the functional ACTN3 p.R577X polymorphism and elite swimmer status in both cohorts were in contrast to many associations with power-/sprint- performance in other sports previously reported. Since current sample size is relatively modest, larger studies will be required to further confirm these results, which, however, have highlighted that it is probable that the genes studied here are not the resulting variants responsible for the phenotypes of interest, despite the associations reported by previous candidate-gene studies in other sports. The present GWAS were conducted in an attempt to identify common polymorphisms associated with elite sprint and endurance status in Jamaicans, African-Americans and Japanese, respectively. These unique athlete cohorts comprised of athletes of the highest standard including world record holders, world champions, Olympians and winners of other international events. Following exclusion of individuals and markers failing the quality control filters, 609,801 autosomal SNPs in 88 Jamaican sprint athletes and 87 Jamaican controls, 637,991 autosomal SNPs in 79 African-American sprint athletes and 391 African-American controls, and 541,179 autosomal SNPs in 114 Japanese athletes (including 60 endurance and 54 sprint athletes) and 116 Japanese controls, were available for association analyses. 17, 7, 36 and 21 SNPs were associated with elite athlete status at a p < 0.00005 threshold of significance in elite Jamaican sprint, African-American sprint, Japanese sprint and Japanese endurance GWAS sets, respectively. Meta-analyses were performed for SNPs with unadjusted association p < 0.00005 across the sprint GWAS sample sets (i.e. Jamaican sprint, African-American sprint, Japanese sprint GWAS cohorts), using the fixed-effects model. The top 17 SNPs (unadjusted p < 0.00005) from the Jamaican sprint cohort were extracted from the association results of African-American sprint, Japanese sprint cohorts, respectively, for the combined effects to be calculated using a meta-analysis method. The same procedure was also applied to the top hits in African-American and Japanese cohorts. The combined odds ratio for the top meta-analysis hit was 2.61 (p = 0.000000466) with the allele G associated with elite sprint status in Jamaicans, African-Americans and Japanese. Although meta-analysis has increased the sample size and power to detect associations in the current GWAS, independent replication of these associations followed by functional studies of replicated SNPs are required. The results of the association studies presented here are the very first positive findings from GWAS involving world-class athletes and these encouraging findings provide further evidence of the importance of genetic predisposition to elite human performance. GWAS of athletes of the highest performance caliber as well as the application of meta-analysis across several initial GWASs seemed to help to circumvent the need for very large cohort of elite athletes and increase the study power. Nevertheless, future GWAS involving large well-funded collaborations using larger cohorts of elite athletes will be necessary in order to explore further the genetic architecture underlying elite human performance. Such initiatives may also allow gene x gene and gene x environment interactions to be explored to some extent, as well as the predictive utility of this genomic research.

Starch digestion begins in the mouth where it is hydrolysed into smaller polysaccharides by the enzyme salivary amylase. Three salivary amylase genes (AMY1A, B & C) and a pseudogene (AMYP1) have been described and are located in tandem on chromosome 1. Polymorphic variation has been demonstrated in Caucasians in the form of the number of repeats of the AMY1 genes, as follows: (lA-lB-Pl)n-lC. This variation has been reported to result in differing levels salivary amylase enzyme production and, as a result, differences in the efficiency of starch digestion in the mouth. It is proposed in this thesis that an increase in salivary gene copy number may be an adaptation to high starch diets as a result of the adoption of agriculture. Reliable high-throughput multiplex PCR based methods have been designed to quantify AMY1 gene copy number and to also to type 6 microsatellite markers closely linked to the AMY gene cluster. Data have been collected for 14 human populations, with different histories of cereal agriculture and ancestral levels of starch in the diet. Data have also been collected on AMY1 gene copy number in 5 chimpanzees (Pan troglodytes). The AMY1 allele frequency difference (measured using FST) between the two most extreme populations, the Mongolians and Saami, was not an outlier on a distribution of FST based on presumed neutral 11,024 SNPs from the human genome. The chimpanzee data suggest that the most frequent allele (AMY1*H1) in humans may not be the ancestral allele, as all chimpanzee chromosomes tested carried the AMY1*H0 allele (containing only one copy of the AMY1 gene). A more sensitive selection test, the analysis of the intra-allelic variability of the AMY1 repeat alleles using closely linked microsatellites, showed no compelling evidence for recent positive selection at the AMY1 locus in humans. As a result, genetic drift could not be ruled out as an explanation for the observed AMY1 allele frequency differences among populations. Alanine:glyoxylate aminotransferase (AGT) is an intermediary metabolic enzyme that is targeted to different organelles in different species. Previous studies have shown that there is a clear relationship between the organellar distribution of AGT and diet. Non-human primates show the herbivorous peroxisomal distribution of AGT. In humans a point mutation and insertion deletion polymorphism have been associated with peroxisome-to-mitochondria AGT mis-targeting. Data have been collected using a PCR/RFLP based method, in 11 human populations. In a comparison with FST values from 11.024 SNP loci, 94.5% of SNPs had a lower FST than a comparison of AGT allele frequencies for Saami and Chinese. This unusually high allele frequency difference between Chinese and Saami is consistent with the signature of recent positive selection driven by the unusually high meat content in the Saami diet.

Environmental signals can directly influence gene expression through epigenetic mechanisms, causing phenotypic changes that can be transmitted to progeny. In plants, this is in part achieved by short interfering RNA (siRNA) which guide covalent modification of DNA, such as cytosine methylation, to specific targets including repetitive sequences and transposable elements. Environmental stress also leads to genome-wide DNA hypomethylation, misregulation of transposable elements and ultimately ‘genomic shock’. Although most stress-induced epigenetic modifications are not thought to be heritable, there is increasing evidence for the inheritance of novel environmentally-induced epigenetic states or ‘environmental epialleles’. The formation of environmental epialleles represents an important source of variation and a powerful driving force of adaptive evolution but the precise mechanism remains unclear. The aim of this thesis is to identify environmental epialleles through computational methods. Analysis of Illumina sequencing data from environmentally stressed maize plants sampled during stress and after a recovery period has so far revealed that a significant proportion of the maize genome is misregulated at both the genetic and epigenetic level. These findings indicate that plants continue to respond after exposure to stress and that this response is likely mediated by at least one epigenetic mechanism, including siRNA-directed DNA modifications.

The circadian clock is the endogenous mechanism by which a wide variety of biological processes are regulated in anticipation of daily changes in the external environment. In Arabidopsis thaliana, the clock comprises a number of complex gene and protein interactions, involving multiple regulatory feedback loops. The clock gene LHY has a central role in these loops, activating and repressing morning- and evening- expressed genes respectively. These clock genes in turn sequentially repress the expression of LHY throughout the day and night, restricting it to a sharp transcriptional peak at dawn. However, the molecular mechanisms of these regulatory interactions with the LHY promoter were not known. Therefore, this project first aimed to determine which promoter motifs are responsible for mediating regulation of LHY circadian expression. This was achieved through luciferase assays with mutated pLHY:LUC reporter constructs, which identified the CT-rich region as responsible for rhythmic expression of LHY, and the G-box as mediating regulation by the clock protein TOC1. Since few regulators were known to target the LHY promoter, this project also aimed to identify transcription factors binding the promoter using a Yeast One-Hybrid assay. Transcription factors with roles in a wide variety of biological pathways were identified from this screen, with abiotic stress and plant defence pathways particularly well-represented. In addition, a number of antagonistic and synergistic regulatory interactions were established as occurring between stress factors and clock proteins at specific promoter motifs. We can therefore conclude that LHY is regulated by a complex network of transcription factor interactions, enabling the rapid integration of environmental stress signals into the clock.

Modern biological research aims to understand when genes are expressed and how certain genes in uence the expression of other genes. For organizing and visualizing gene expression activity gene regulatory networks are used. The architecture of these networks holds great importance, as they enable us to identify inconsistencies between hypotheses and observations, and to predict the behavior of biological processes in yet untested conditions. Data from gene expression measurements are used to construct gene regulatory networks. Along with the advance of high-throughput technologies for measuring gene expression statistical methods to predict regulatory networks have also been evolving. This thesis presents a computational framework based on a Bayesian modeling technique using state space models (SSM) for the inference of gene regulatory networks from time-series measurements. A linear SSM consists of observation and hidden state equations. The hidden variables can unfold effects that cannot be directly measured in an experiment, such as missing gene expression. We have used a Bayesian MCMC approach based on Gibbs sampling for the inference of parameters. However the task of determining the dimension of the hidden state space variables remains crucial for the accuracy of network inference. For this we have used the Bayesian evidence (or marginal likelihood) as a yardstick. In addition, the Bayesian approach also provides the possibility of incorporating prior information, based on literature knowledge. We compare marginal likelihoods calculated from the Gibbs sampler output to the lower bound calculated by a variational approximation. Before using the algorithm for the analysis of real biological experimental datasets we perform validation tests using numerical experiments based on simulated time series datasets generated by in-silico networks. The robustness of our algorithm can be measured by its ability to recapture the input data and generating networks using the inferred parameters. Our developed algorithm, GBSSM, was used to infer a gene network using E. coli data sets from the different stress conditions of temperature shift and acid stress. The resulting model for the gene expression response under temperature shift captures the e�ects of global transcription factors, such as fnr that control the regulation of hundreds of other genes. Interestingly, we also observe the stress-inducible membrane protein OsmC regulating transcriptional activity involved in the adaptation mechanism under both temperature shift and acid stress conditions. In the case of acid stress, integration of metabolomic and transcriptome data suggests that the observed rapid decrease in the concentration of glycine betaine is the result of the activation of osmoregulators which may play a key role in acid stress adaptation.

The complexity and size of the higher animal genome and relative scarcity of DNA-binding factors with which to regulate it imply a complex and pleiotropic regulatory system. Cisregulatory modules (CRMs) are vitally important regulators of gene expression in higher animal cells, integrating external and internal information to determine an appropriate response in terms of gene expression by means of direct and indirect interactions with the transcriptional machinery. The interaction space available within systems of multiple CRMs, each containing several sites where one or more factors could be bound is huge. Current methods of investigation involve the removal of individual sites or factors and measuring the resulting effect on gene expression. The effects of investigations of this type may be masked by the functional redundancy present in some of these regulatory systems as a result of their evolutionary development. The investigation of CRM function is limited by a lack of technology to generate and analyse combinatorial mutation libraries of CRMs, where putative transcription factor binding sites are mutated in various combinations to achieve a holistic view of how the factors binding to those sites cooperate to bring about CRM function. The principle work of this thesis is the generation of such a library. This thesis presents the development of microstereolithography as a method for making microfluidic devices, both directly and indirectly. A microfluidic device was fabricated that was used to generate oligonucleotide mixtures necessary to synthesise combinatorial mutants of a CRM sequence from the muscle regulatory factor MyoD. In addition, this thesis presents the development of the optimisation algorithms and assembly processes necessary for successful sequence assembly. Furthermore, it was found that the CRM, in combination with other CRMs, is able to synergistically regulate gene expression in a position and orientation independent manner in three separate contexts. Finally, by testing a small portion of the available combinatorial mutant library it was shown that mutation of individual binding sites within of the CRM is not sufficient to show a significant change in the level of reporter gene expression.

Colletotrichum acutatum sensu lato includes a number of important pathogens that cause economically significant losses of various crops. The C. acutatum species complex has a wide host range in both domesticated and wild plant species, and its capability to infect different types of hosts such as insects has also been described. Members belonging to this complex are able to develop three different types of interaction with plant hosts including biotrophic, necrotrophic and hemibiotrophic infections and are also capable of surviving on weeds and non-hosts without causing visible symptoms. They are mainly asexual, but some have a teleomorphic state called Glomerella and can be either homothallic or heterothallic. The sexual behaviour in Glomerella is more complicated than in most ascomycetes, and strains within the same species do not show a typical MAT1-1/2 system. The overall aim of this study was to gain an improved understanding of the relationships between the genetic diversity of global populations, host association patterns, geographic distribution and biological and pathological attributes. A database (CaITSdb) containing more than 800 rRNA sequences deposited in GenBank was created along with key biogeographic information, and the data have been analysed in order to investigate genetic diversity and distribution of sub-populations and their evolutionary relationships. The combined evidence was used to assemble a core collection of 120 isolates that are representative of the diversity in host preference, geographic origin, mating behaviour and molecular genetic variation. A multi-locus sequencing approach (based on four neutral loci) has been used to evaluate phylogenetic relationships amongst the isolates in the core collection. A strong relationship was observed between various genetic groups distinguished and their mating behaviour, geographic distribution and host association patterns. Oceania has been identified as a likely geographic origin of this pathogen, as the highest level of variability and groups related to a hypothetical ancestral population are mainly distributed in these countries. All homothallic isolates capable of producing perithecia belonged to the same genetic group A7; whereas all self-sterile heterothallic isolates were classified as either A3 or A5. Isolates derived from the same host tend to cluster together into genetic groups or sub-populations. This evidence is generally in agreement with recent published work on taxonomic re-assessment of Colletotrichum acutatum sensu lato, which indicates at least fifteen new species. This study has provided the evidence for the occurrence of three distinct genetic groups on strawberry in the UK corresponding to three species reported in the literature namely, C. nymphaeae, C. fioriniae and C. godetiae. Isolates belonging to the genetic groups that correspond to C. nymphaeae and C. fioriniae appeared to be the most aggressive on strawberry, followed by C. godetiae, and C. simmondsii (not found in the UK). Representative isolates of other species were less aggressive. The first whole genome sequence an isolate (A9 = C. simmondsii) from the C. acutatum sensu species complex was assembled and analysed using a range of bioinformatics algorithms. An isolate of C. simmondsii was chosen based on its wide host range including strawberry and the phylogeographic position. Genome analyses enabled prediction and annotation of the whole gene set at 13549 including 6 % unique to this species. The data also suggested an interesting expansion of several gene families, such as those encoding carbohydrate-active enzymes, secondary metabolites pathways and effectors which could be associated with the wide host range. The new knowledge and resources developed with the genome analyses along with the results of the population level diversity studies provide a platform for future comparative and functional genomics investigations to advance this research.

Two conserved protein kinases, Chk1 and Chk2, are activated in response to genotoxic stress and mediate multiple cell cycle checkpoint mechanisms that ensure genomic integrity. The establishment of mitotic checkpoint delay in response to DNA damage or incompletely replicated DNA is conventionally thought to be accomplished through inhibition of the cyclin-dependent kinase, Cdc2. Both Chk1 and Chk2 have the potential to operate in this pathway. Therefore, the initial aim of this study was to investigate the relative requirement of Chk1 and Chk2 for mitotic delay mechanisms triggered by DNA damage and DNA replication arrest in avian and human cells. These studies demonstrated that Chk1 is the principal regulator of the G2/M checkpoint with a direct role in the establishment and maintenance of the mitotic delay in response to DNA damage. Chk1 was also found to be required for the S/M checkpoint in response to DNA replication arrest; however, detailed analysis indicated that its role is to maintain rather than initiate this checkpoint, as cells lacking functional Chk1 can initially delay mitosis for many hours before they enter a premature mitosis with unreplicated DNA. In avian cells, mitotic phosphorylation of cyclinB2 was found to be mediated by cyclin dependent kinases and suppressed by checkpoint signalling. However, accumulation of potentially active phospho-cyclinB2/Cdc2 complexes was observed during the initial mitotic delay in the absence of functional Chk1, suggesting that other factors apart from the conventionally known mechanisms can restrain mitotic Cdc2 activity. In addition, avian cells were able to delay mitosis effectively during replication arrest in the presence of the ATM/ATR inhibitor caffeine, further emphasizing the possibility of mitotic delay mechanisms that operate independently of ATM/ATR and Chk1. Furthermore, this study revealed that endogenous Cdc6 accumulates in a Chk1-dependent manner during replication arrest. To test whether Cdc6 might function upstream or downstream of Chk1 in the replication checkpoint pathway, Cdc6 was ectopically expressed in both checkpoint-proficient and checkpoint-deficient Chk1-depleted cells. The results from these intervention experiments give preliminary evidence that places Cdc6 downstream of Chk1 in the S/M checkpoint response. The ability of cells to delay the onset of mitosis while DNA replication is stalled independently of ATM/ATR/Chk1 is consistent with the general idea of an inherent relationship between the process of DNA replication and mitosis. The replication machinery might be able to signal either normal DNA replication in progress or the presence of stalled replication structures and thereby intrinsically link the successful completion of DNA synthesis to progression into mitosis.