The Family Study of Myopia is a research project aiming to discover genetic loci causing susceptibility to high myopia. As part of this investigation, the heritability of refractive error and other ocular components was estimated for a large Irish-Welsh multi-generational pedigree using variance components analysis software, SOLAR. Heritabilities of 0.39 (p=6.92 x 10"5, S.E.=0.14), 1.00 (p=1.84 x 10"4, S.E.=0.22) and 0.30 (p=0.13, S.E.=0.33) were found for refractive error, mean corneal curvature and axial length, respectively. Heritability of refractive error was also calculated using within-family regression and a Markov Chain Monte Carlo method producing estimates of between 0.12 and 0.73. This high heritability of ocular refraction suggests the potential for finding susceptibility loci for the control and development of refractive error. To pinpoint these loci, areas of chromosomes which have previously been suggested to harbour genes controlling refractive error were investigated to determine whether linkage was present within this family. DNA was extracted from mouthwashes and linkage analysis was performed. SOLAR revealed no significant linkage to the loci tested, reinforcing the theory that myopia is a highly heterogenous disease. The maximum twopoint LOD score was within the MYP6 locus at marker D22S1176 (LOD=1.19). Multipoint analysis showed the maximum LOD score at the MYP3 locus, between markers D12S1605 and D12S354 (LOD =1.37). In a separate study of 96 families containing a highly myopic child and the two parents, the involvement of a candidate gene encoding the protein myocilin (MYOC) was examined using an association analysis. There was weak evidence of over- transmission of allele 3 of MYOC 1, a marker in the 5' untranslated region of the gene, and under-transmission of allele 4 of that same marker (both p<0.05). This suggests MYOC may play a small role in the causation of high myopia development but a larger sample is needed to establish this conclusively.

Since Kraepelin first dichotomised the functional psychoses into dementia praecox and manic-depressive insanity at the end of the 19th century, the validity of the distinction has been challenged. Phenomenological, neurobiological, family, and molecular-genetic studies suggest that there is no neat biological distinction between these entities which are now known as schizophrenia and bipolar disorder. The aim of this thesis was to explore the familial correlation of clinical measures within a large harmonised clinical dataset comprising samples of (a) families enriched for bipolar disorder, and (b) families enriched for schizophrenia. Analyses were performed across traditional diagnostic boundaries. I carried out systematic clinical ratings on 835 individuals previously collected as part of ongoing molecular genetic studies. After an intensive training period, which included reliability exercises, I rated each case on approximately 200 variables, including a new set of rating scales developed as part of the PhD project. I performed mixed-effects regression analysis on the data to estimate the intra-class correlation coefficient (ICC) and significance for each variable. After controlling for sample-of-origin and gender, thirty-one variables were significantly correlated within families. Amongst the most significant were age at onset (ICC=0.287, p=0.0006), longest admission (ICC=0.287, p=0.0006) and cannabis abuse/dependence (ICC=0.639, p=0.0007). Such variables may be influenced by genetic factors and may therefore be used to identify subgroups of patients more likely to share common underlying genetic susceptibilities. In an analysis of a subset of sibling-pairs that were enriched for schizoaffective disorder I found that genetic similarity at chromosome lq42 was significantly associated with phenotypic similarity for the most severe depressive episode. I also undertook clinical ratings on a sample of previously-collected patients with Velo-Cardio-Facial Syndrome (VCFS) and found high-rates of both mood-disturbance and psychosis. My findings show that clinical ratings can be a useful adjunct to categorical diagnoses and identify specific phenotypes to consider in genetic studies.

The overall aim of this study is to evaluate a diverse selection of methods for the analysis of large-scale gene expression data derived from human brain, and to apply them to furthering the understanding of heritable psychiatric disorders. One strand of research presented here focuses on using clustering algorithms to group genes according to their expression. Several methods for expression clustering were implemented and based upon brain expression datasets. Gene Ontology enrichment was then used to assess the concordance of the resulting clusters with current biological knowledge. Combining different clustering methods is the most effective strategy, as it allows the discovery of the widest range of clusters. Clusters produced by these methods were then investigated for enrichment with genes associated with, or differentially expressed in, bipolar disorder or schizophrenia. Particularly enriched clusters were further studied using the functional annotation database MetaCore. The second strand of this research focused on using control adult brain expression data and expression quantitative trait analysis to divide SNPs into those with a greater and lesser effect on global gene expression. This classification was used to enhance the prediction of schizophrenia affected status from genome-wide association study SNP data using polygenic score analysis, a method which aggregates information from a large number of loci. SNPs which have a larger effect on global gene expression are significantly superior at predicting schizophrenia affected status through polygenic score analysis, a novel finding which suggests that expression data from control adult brain can have relevance to the study of schizophrenia.

In order to investigate the role of nonsynonymous variants of the APC gene in inherited predisposition to colorectal adenomas (CRAs) the entire APC ORF was sequenced in 691 unrelated North American patients with CRAs and 969 healthy controls. There was significant over representation of rare inherited nonsynonymous variants in patients who did not carry conventional pathogenic mutations in APC or MUTYH (P = 0.0113) when compared to patients with familial adenomatous polyposis (FAP) and MUTYH associated polyposis (MAP). The over representation was highest in non-FAP non-MAP patients with 11 to 99 CRAs (P = 0.0103). More non-FAP non-MAP patients carried rare nonsynonymous variants within the functionally important p-catenin down-regulating domain compared with healthy controls (P = 0.0166). In silico analyses predicted that approximately 46% of the variants identified were expected to affect function. Functional characterisation in vitro showed that 7 of 16 nonsynonymous variants altered p-catenin-regulated transcription consistent with a role in predisposition to CRAs. An optimum level of p-catenin signalling is proposed to drive colorectal tumourigenesis, mediated by selection for APC genotypes retaining one, or rarely two, 20 amino acid p-catenin down regulating repeats (20AARs). We investigated the mechanism through which the APC variant E1317Q contributes to colorectal tumourigenesis. We compared the somatic mutation spectra of APC in tumours from AFAP patients that did (Family B) or did not (Family S) co-inherit E1317Q. Significant differences were identified between these tumours, 8.2% of tumours carrying E1317Q had somatic mutations predicted to result in mutant polypeptides retaining a single 20AAR, as compared to 62.1% of those which did not carry this variant (P=5.64x10 9). In vitro assays showed that E1317Q significantly impaired p-catenin regulated transcription when expressed with 'weak' truncating mutations (P<0.05) suggesting that E1317Q relaxes the target for tumourigenic somatic APC mutations through its own effects on p-catenin-associated signalling. Inherited mutations in the MUTYH gene predispose to an autosomal recessive disorder characterise by multiple CRAs and carcinomas (MAP). MUTYH is a DNA glycosylase which removes adenines that are misincorporated opposite 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxoG), one of the most stable products of oxidative DNA damage. Tumours from patients with MAP display a high proportion of somatic G:C->T:A mutations due to a failure to repair these mismatches, and it is presumed that this mutator phenotype drives tumourigenesis. We studied the response of primary MUTYH-deficient fibroblasts to oxidative stress and found that significantly more of these cells survived exposure to hydrogen peroxide and f-butyl-hydroperoxide as compared to wild type cells. We found that MUTYH-deficient cells failed to enter apoptosis and showed that this may be mediated via an inability to recruit Rad9 to the correct nuclear position, indicating a failure to engage the 9-1-1 DNA damage sensor complex. Consistent with this, we found that MUTYH-deficient cells failed to activate the downstream checkpoint protein Chk1, after exposure to oxidative stress. We propose that MAP-associated tumourigenesis is driven by failure to undergo apoptosis in conjunction with an underlying mutator phenotype.

The search for genetic variants that alter the risk of developing schizophrenia has met with little success (Owen et al., 2005a) with the best example coming from a large multiply affected pedigree (Blackwood et al., 2001, Millar et al., 2000, St Clair et al., 1990). In this study a genome-wide significant schizophrenia linkage region in a single pedigree (Williams et al., 2003a) has been refined to an 11.7Mb region at 17q23-q24 where all 6 affected males share 21 consecutive microsatellite marker genotypes. Analysis of this region by oligonucleotide-array comparative genome hybridisation shows that no large deletions explain the linkage signal. High-density genotyping identified a region of homozygosity present in C702 affecteds that was not identified in 2709 individuals from the UK. This rare diplotype encompasses the 3' of the gene encoding Protein Kinase C Alpha (PRKCA), implicated in the pathogenesis of schizophrenia and related disorders (Mimics et al., 2001, Hahn and Friedman, 1999, Birnbaum et al., 2004). Mutation screening of PRKCA in the linked pedigree C702 identified an exonic haplotype with a minor allele frequency of 0.003, which is homozygous in affected individuals. The haplotype is associated in a UK case-control sample with major mental illness (p=0.05. OR=l .8) due to risk in males (p=0.005. OR=3.9). Also, the pedigree C702 genotype was not observed in -9000 Europeans. Association mapping of PRKCA using schizophrenia and psychosis case-control samples from Europe identified a common associated allele (rs3803821, meta analysis p=0.02, OR=l.l) that shows significant overtransmission in a trios sample (p=0.03). Therefore, PRKCA may represent the locus causing the pedigree C702 linkage signal and contains genetic variation associated with schizophrenia and related disorders.

Dysbindin (dystrobrevin-binding protein 1 DTNBP1) has been implicated as a schizophrenia candidate gene. However while numerous positive associations have been reported, no non-synonymous alleles have been found which account for the association. A number of recent studies suggest that altered dysbindin expression may be the mechanism by which DTNBP1 variants confer susceptibility to schizophrenia. Therefore one objective of this study was to identify putative DTNBP1 cis-acting variants and perform association analyses of these variants with schizophrenia and allelic expression differences observed at the DTNBP1 locus. While four variants were associated with schizophrenia, logistic regression suggested that the signal observed at these polymorphisms was not independent of the most associated SNP rs4715984. Comparison with the results of a DTNBP1 allelic expression assay revealed that seven SNPs were associated with differential expression. Post hoc analysis revealed that the majority of the expression differences were accounted for by variation at two loci (rs2619538 and rsl3198512), one of which (rsl3198512) was subsequently shown to directly affect transcription in vitro using a luciferase reporter gene assay. As rs4715984 was not correlated with allelic expression differences it implies that a reduction in dysbindin expression through cis-acting variation may not be the primary aetiological factor in schizophrenia pathogenesis. This was supported by further analysis of a schizophrenia risk haplotype previously reported to be associated with differential expression as the refined haplotype was no longer correlated. A second objective of this thesis was to investigate the hypothesis that DTNBP1 could cause susceptibility to schizophrenia through its role within the BLOC-1 complex. Association analysis was performed on BLOC-1 genes which displayed evidence of being under the influence of cis-acting regulation. MUTED, a BLOC-1 gene previously reported as associated to schizophrenia was also investigated. However association results provided no compelling support for the hypothesis that DTNBP1 contributes to susceptibility to schizophrenia through the BLOC-1 complex.

Darier's Disease (DD) is a rare autosomal dominantly inherited skin disorder in which co-occurrence of neuropsychiatric abnormalities has been frequently reported by dermatologists. The disease is caused by mutations in a single gene, ATP2A2, which maps to 12q23-q24.1 and is expressed in the skin and brain. This gene encodes SERCA2 (sarco/endoplasmic reticulum Ca 2+ ATPase isoform 2), a calcium pump involved in intracellular calcium transport. This study aimed to conduct the first systematic investigation of the neuropsychiatric phenotype in DD, investigate possible genotype-phenotype correlations, and compare the neuropsychiatric features in DD individuals and their first-degree unaffected relatives. One hundred unrelated individuals with DD and 24 of their unaffected relatives were assessed using a battery of standardised neuropsychiatric measures. The relationship between the mutations detected in iheATP2A2 gene and the presence/severity of neuropsychiatric phenotypes was examined. DD individuals reported high rates of mood disorders, specifically major depression (30%), suicide attempts (13%) and suicidal thoughts (31%), and these were significantly more common in DD when compared with normative population data. Further, individuals with DD reported higher scores on measures of neuropsychiatric dysfunction than their unaffected relatives. These associations cannot be explained by psychosocial factors. Mutations found among individuals with similar neuropsychiatric phenotypes clustered in certain locations within the SERCA2b protein. Together, these findings support the hypothesis that mutations mATP2A2 confer susceptibility to neuropsychiatric dysfunction, in particular mood disorders, in individuals with DD. These findings highlight the need for assessment and recognition of psychiatric symptoms in DD. The findings may also have implications for identification of other genetic factors involved in conferring susceptibility to neuropsychiatric features in individuals without DD. Further research is needed into other neuropsychiatric phenotypes in DD and into the specific functional effects of mutations in ATP2A2 and the relationship of these to the presence of certain neuropsychiatric phenotypes.

In recent years, molecular genetics research has identified a large number of putative susceptibility genes for a variety of complex psychiatric and neurodegenerative disorders. However, in most instances, the particular functional variants involved have not been identified, and it is typically unclear by what mechanism the pathogenic effect is mediated. Where a genetic association does not appear to be fully explicable by variants that alter the amino acid sequence of a protein, it is a reasonable hypothesis that the association might be mediated by cis-acting variants that alter gene expression. This hypothesis was tested in this thesis in relation to 10 putative susceptibility genes for psychiatric and neurodegenerative disorders. The genes were DISCI, RELN, GABRA4, GABRA5, GABRB1, GABRB2, GABRG2, GABRG3, NOS1AP and MAPT. Each one of these genes was investigated by assays of relative allelic expression applied to a large number of post mortem human brain samples. Samples were also genotyped for relevant variants that had previously shown association with disease in order to test those variants for a putative cis-regulatory effect. Cis-regulatory variation manifested as unequal expression of each parental gene copy at the mRNA level was detected in nearly all of the genes in at least one tissue sample. However, for only two genes (RELN and MAPT) was evidence obtained that specific variants implicated in disease influenced expression.

The known somatic (N>4000) and germline (N>4000) cancer-associated mutational spectra (viz. missense and nonsense mutations micro-deletions, micro-insertions and micro- indels <20bp) of 17 human tumour suppressor genes (viz. APC, ATM, BRCA1, BRCA2, CDH1, CDKN2A, NF1, NF2, PTCH, PTEN, RBI, STK11, TP53, TSC1, TSC2, VHL and WT1) were compared in order to identify similarities and differences. Analysed parameters included the recurrence status of mutations, CpG mutability Grantham difference evolutionary conservation of affected codons role of nonsense-mediated mRNA decay and co-location with repetitive sequence elements. Only a small proportion of the mutations (-5%) were found to be shared between the germline and soma, although the proportions varied between different types of mutation (from 11% for missense mutations to 1% for micro-indels). Shared mutations are unlikely to be coincidental and are probably indicative of underlying shared (and endogenous) mutational mechanisms. Shared missense mutations were found to be more likely to be drivers of tumorigenesis than either exclusively somatic or exclusively germline missense mutations. Shared micro-lesions combined for all genes occurred disproportionately within repetitive elements by comparison with both somatic or germline micro-lesions, consistent with an endogenous mutational mechanism. For some genes (e.g. TP53), shared CpG-dinucleotide mutations evidenced the action of an endogenous mutational mechanism (viz. methylation-mediated deamination of 5-methylcytosine) in both the soma and the germline. Differences between mutational spectra were also noted. Germline missense mutations were found to be more likely to bear relatively more drastic functional consequences by comparison with somatic missense mutations, but also more likely to be truncating mutations. Germline micro-lesions (combined for all genes) were also found to be more likely to be co-located with repetitive elements than somatic micro-lesions. This could be due to the germline being relatively more protected from the action of exogenous mutagens by comparison to the soma. This study of 17 human tumour suppressor genes has therefore provided a first glimpse of the similarities and differences between germline and somatic mutational spectra.

NER is vital for the integrity of the genome, with defects in NER giving rise to the rare human disorder Xeroderma Pigmentosum. This study has seen the construction of a minichromosome containing MFA2 as a model gene. The TAM plasmid model provides a chromatin environment analogous to that seen at the endogenous MFA2 gene, and the nucleosome positions have been determined along with the repair profile in wild type cells, and in cells lacking the general repressor Tup1p. When TUP1 is deleted, repair of CPDs occurs at a faster rate in both the TS and NTS of the MFA2 promoter. Repair was studied in respect to the HAT Gcn5p, a factor responsible for the H3 acetylation (K9 and K14) of MFA2, a prerequisite for efficient transcription initiation and repair. The TAM model showed diminished repair and a reduction in H3 acetylation at the MFA2 gene when GCN5 was absent, confirming the role for Gcn5p in repair at this gene, regardless of the chromosomal context. Repair was also investigated in the absence of the GG-NER factor Rad16p. Rad16p has been implicated with roles in chromatin remodelling, as well as the UV-induced occupancy of Gcn5p to promote H3 acetylation. In TAM, the absolute requisite for Rad16p in GG-NER was overcome, suggesting the chromatin environment within the plasmid differs from that seen at the endogenous MFA2. The TAM plasmid has provided a tool to study NER both <italic>in vivo,</italic> as reported here, and <italic> in vitro,</italic> enabled the biochemistry behind the complex mechanism of NER in chromatin to be realised.