A major challenge in human genetics is to understand the genetic variation underlying common diseases. In this thesis, I focus on forces creating differences between individuals and genomic regions, methods for characterizing genomic variation, and the association between genomic and phenotypic variation. Genetic markers are widely used to locate genes associated with different phenotypes. In my first paper, I describe novel algorithms for automatic genotype determination of microsatellite markers, a procedure which is currently both time-consuming and error prone. The co-segregation of genetic markers in a population leads to non-random association of alleles at different loci - linkage disequilibrium (LD). LD varies throughout the genome and differs between populations due to factors such as their demographic history. In my second paper, I discuss the increased power, for mapping of human traits, that results from studying a population with appreciable levels of LD such as is found in the Swedish Sami population. Lately, large-scale analyses of single nucleotide polymorphisms (SNPs) have become available and efforts have been made to identify a set of SNPs, which captures most of the genome variation in a population (tagSNPs). In my third paper, I describe the limitations of this approach when applied to data from an independent population sample of randomly ascertained SNPs. The transferability of tagSNPs between populations is poor, presumably due to variation in allele frequencies and the bias towards common SNPs used in most studies. The level of genomic variation is influenced by population structure, recombination and mutation rate, as well as natural selection. During the exodus from Africa, humans have adapted to new environmental conditions. In my fourth paper, I describe a new method for identifying genomic regions carrying signatures of recent positive selection and apply this to an available dataset of millions of SNPs.

Many problems in bioinformatics are inference problems, that is, the problem objective is to infer something based upon a limited amount of information. In this work we explore two different inference problems in bioinformatics. The first problem is inferring the structure of signal transduction networks from interactions between pairs of cellular components. We present two contributions towards the solution to this problem: an mixed integer program that produces and exact solution, and an implementation of an approximation algorithm in Java that was originally described by DasGupta et al. An exact solution is obtained for a problem instance consisting of real data. The second problem this thesis examines is the problem of inferring complete haplotypes from informative SNPs. In this work we describe two variations of the linear algebraic method for haplotype prediction and tag SNP selection: Two different variants of the algorithm are described and implemented, and the results summarized.

The search for genetic regions associated with complex diseases, such as cancer or Alzheimer's disease, is an important challenge that may lead to better diagnosis and treatment. The existence of millions of DNA variations, primarily single nucleotide polymorphisms (SNPs), may allow the fine dissection of such associations. However, studies seeking disease association are limited by the cost of genotyping SNPs. Therefore, it is essential to find a small subset of informative SNPs (tag SNPs) that may be used as good representatives of the rest of the SNPs. Several informative SNP selection methods have been developed. Our experiments compare favorably to all the prediction and statistical methods by selecting the least number of informative SNPs. We proposed algorithms for faster prediction which yielded acceptable trade off. We validated our results using the k-fold test and its many variations.

In one study, a suite of plants from the maize genus Zea L. (Poaceae) and the specialist herbivore Dalbulus maidis (DeLong and Wolcott, 1923) (Hemiptera: Cicadellidae) were used to test the hypotheses that anti-herbivore defenses are affected by plant life-history evolution and human intervention through domestication and breeding for high yield. The suite of Zea host plants included one Mexican commercial hybrid maize Zea mays ssp. mays L., a landrace variety of maize, two populations of Balsas teosinte (Zea mays ssp. parviglumis Iltis & Doebley), and perennial teosinte (Z. diploperennis Iltis, Doebley & Guzman). This suite of host plants includes three Transitions evident within the genus Zea: life history form perennial to annual life cycle evident between perennial teosinte and Balsas teosinte, domestication transition from wild annual to domesticated annual evident between Balsas teosinte and landrace maize, and; breeding transition from landrace cultivar to a hybrid cultivar. The transitions were correlated with differences in plant defenses, as indicated by corn leafhopper performance. Results showed a performance gradient, suggesting a pattern in which plant defense is stronger in perennial than annual plants, Balsas teosinte than landrace maize, and in landraces than in hybrid maize. Furthermore, results suggested that domesticated maize would be the least defended, most suitable host for corn leafhopper. In a second study, haplotype diversity was assessed to address structuring and interconnectedness among samples of corn leafhopper collected in the southwestern region of Mexico to address microevolution. The geographic focus of the study was maintained within an area encompassing the presumed centers of radiation of Dalbulus and its host genus Zea, and of maize domestication. Samples were complemented with samples of corn leafhopper sequences available at GenBank. Results revealed seven haplotypes from three host plants within Zea: perennial teosinte, Balsas teosinte, and maize. Furthermore, genetic differentiation was present and haplotype diversity appears to correlate with differences in genetic structure between perennial teosinte and maize. One haplotype was found to be present throughout all sites, which appears to parallel the spread of maize cultivation. As maize cultivation spread beyond its area of domestication, corn leafhoppers colonized perennial teosinte, further suggesting that subsequent decreases in maize cultivation in perennial teosinte habitat created a refuge where perennial teosinte- adapted haplotypes could persist. Altogether, my research suggests that the combination of historical expansion of maize cultivation expansion and the weaker anti-herbivore defenses associated within maize domestication appears to have favored genotypes particularly adapted for exploiting maize.

The Human Genome Project completed in 2003 and the draft of human genome sequences were also yielded. It has been known that any two human gnomes are almost identical, and only very little difference makes human diversities. Single nucleotide polymorphism (SNP) means that a single-base nucleotide changes in DNA. A SNP sequence from one of a pair of chromosomes is called a haplotype. In this thesis, we study how to reconstruct a pair of chromosomes from a given set of fragments obtained by DNA sequencing in an individual. We define a new problem, the chromosome pair assembly problem, for the chromosome reconstruction. The goal of the problem is to find a pair of sequences such that the pair of output sequences have the minimum mismatch with the input fragments and their lengths are minimum. We first transform the problem instance into a directed multigraph. And then we propose an efficient algorithm to solve the problem. We apply the ACO algorithm to optimize the ordering of input fragments and use dynamic programming to determine SNP sites. After the chromosome pair is reconstructed, the two haplotypes can also be determined. We perform our algorithm on some artificial test data. The experiments show that our results are near the optimal solutions of the test data.

Sequence variation of a 238 bp segment of the mitochondrial control region was analyzed for 1,568 Steller sea lions (2.8% of the estimated species population) sampled from 50 rookeries representing nearly every locality at which Steller sea lions are known to breed in significant numbers. Haplotype diversity (H = 0.9164 ± 0.0035) was high and nucleotide diversity (π = 0.00967 ± 0.00586) was moderate. No evidence was observed for significant genetic bottleneck effects. Rookeries were grouped into regions and stocks to examine structure at different spatial scales. F- and Φ-statistics were computed for all pairwise comparisons of rookeries, regions and stocks. Significant (P<0.05) divergence of eastern stock (southeastern Alaska to California) animals from western stock animals was supported in analyses at all spatial scales. Likewise, rookeries and regions from Asia were found to be significantly different from all other western stock rookeries. This was most clearly demonstrated using Φ-statistics at the regional level. The Commander Islands clearly associate with Alaskan western stock rookeries, not with the Asian rookeries. Within each of the three stocks there is significant isolation by distance among rookeries. This relationship does not hold for inter-stock comparisons indicating that there are important barriers to gene flow among stocks. Mitochondrial DNA analysis supports the recognition of three stocks for appropriate conservation of the species. The currently recognized eastern stock is unaffected, but the western stock is now partitioned west of the Commander Islands yielding a western stock which ranges from Prince William Sound west to the Commander Islands, and an Asian stock including rookeries from the Kamchatka Peninsula, Kuril Islands, and Sea of Okhtosk.

Obesity is a complex and a highly individualized disease and the molecular mechanisms behind this disorder need to be better elucidated. Identification of genes and genetic variants that are involved provide opportunities to establish a genetic understanding of the disease. These findings may also provide more rational approaches to therapy, either by identifying underlying causes or point out the need for different treatments. In addition, the timing and severity of obesity may provide insights into the aetiology of obesity and also identify age-specific determinants of weight gain. Recently, genome-wide association studies have led to a rapid progress in our understanding of the genetic basis of various diseases and candidate genes for obesity have been identified. The overall aim of this thesis was to investigate the genetic impact on severity of childhood obesity and the associations between obesity and genetic variants in the fat mass and obesity associated gene, FTO, and MGAT1, the gene encoding mannosyl (α-1,3-)-glycoprotein β-1,2-N-acetyl-glucosaminyltransferase. We show that the impact of parental body mass index (BMI) on the severity of obesity in children is strengthened as the child grows older, whereas the age at obesity onset is of limited importance. By association studies, we show that single nucleotide polymorphisms downstream MGAT1 influence susceptibility to obesity. Moreover, these variants affect the levels of unsaturated fatty acids and desaturase indices, variables previously shown to correlate with obesity. Furthermore, one variant in the first intronic region of FTO is associated with obesity among children but not with BMI or other measures of adiposity at older ages. However, this variant shows a weight-dependent association with cognitive function among elderly men. By direct sequencing, we identified novel variants in FTO, affecting glucose homeostasis in a BMI-independent manner. Furthermore, we found gender specific effects for FTO, both regarding obesity susceptibility and related phenotypes.

The metabolism of vitamin D commences with exposure of the skin to sunlight. The growing recognition of its role in insulin resistance, autoimmune disorders, infections, cancer, as well as the health of cells that influence physical and mental function have profound implications on how we define vitamin D requirements and why we should care whether they are met or not. Most of the actions of vitamin D are mediated by the vitamin D receptor (VDR), a protein whose gene sequence can vary, giving rise to polymorphic forms which are potent enough to affect the binding capacity of this protein to vitamin D. Some of these polymorphic forms of VDR gene may be associated with reduced effectiveness of vitamin D and hence predispose individuals to diseases such as type 2 diabetes and insulin resistance. An earlier study, the Surya Study, looked at the responsiveness of the South-Asian women living in Auckland to vitamin D. The research described here is an extension of this study and its focus was to identify the associations/linkages between certain polymorphic forms of the VDR gene and the disease conditions and intervention responsiveness in the same women. The first objective was to compare two well known techniques for genotyping single nucleotide polymorphisms (SNPs) of the VDR gene at the 3’ end, namely BsmI, ApaI and TaqI: the newer real-time polymerase chain reaction (qPCR) and the traditional restriction fragment length polymorphism polymerase chain reaction (RFLP-PCR) techniques. This comparison was performed to evaluate alternative methods for genotyping which consumed less time than RFLP-PCR. When the presence of each polymorphism by both the techniques was compared in this cohort of South-Asian women, it was found that RFLP-PCR proved to be a more reliable technique than qPCR for genotyping the VDR gene. Another objective of this project was to investigate the prevalence of the above three polymorphisms along with Cdx-2 and FokI SNPs which are present at the 5’ end of the VDR gene, in the population under study and their possible association with phenotypes such as vitamin D responsiveness and insulin resistance. These women were screened and biochemical data was collected during the earlier Surya Study. Of these, eighty-one women were then selected for intervention based on them having high insulin resistance (HOMA-IR>1.93) and serum 25(OH)D<50 nmol/L. Out of these eighty-one women, forty-two were given vitamin D supplement and thirty-nine were given a placebo for six months. Baseline and endpoint measurements included insulin resistance (HOMA-IR), insulin sensitivity (HOMA2%S) etc. How each individual responded to treatment in the intervention group was analysed in the context of the polymorphisms that they had. An association of insulin resistance with BsmI, ApaI and TaqI SNPs was observed in this cohort of 239 women. The response to insulin resistance in the vitamin D supplemented group significantly differed for FokI genotype compared to other genotypes. This explained why certain women responded to treatment better than the others. When the frequencies of the genotypes of these five SNPs of the VDR gene were compared to other studies of different ethnicities, the results of this study were consistent with few studies but contradictory to others. The possible reasons for these differences could be because of small sample size and different ethnicities under study due to which the frequency of alleles and hence the genotypes differed.

The metabolism of vitamin D commences with exposure of the skin to sunlight. The growing recognition of its role in insulin resistance, autoimmune disorders, infections, cancer, as well as the health of cells that influence physical and mental function have profound implications on how we define vitamin D requirements and why we should care whether they are met or not. Most of the actions of vitamin D are mediated by the vitamin D receptor (VDR), a protein whose gene sequence can vary, giving rise to polymorphic forms which are potent enough to affect the binding capacity of this protein to vitamin D. Some of these polymorphic forms of VDR gene may be associated with reduced effectiveness of vitamin D and hence predispose individuals to diseases such as type 2 diabetes and insulin resistance. An earlier study, the Surya Study, looked at the responsiveness of the South-Asian women living in Auckland to vitamin D. The research described here is an extension of this study and its focus was to identify the associations/linkages between certain polymorphic forms of the VDR gene and the disease conditions and intervention responsiveness in the same women. The first objective was to compare two well known techniques for genotyping single nucleotide polymorphisms (SNPs) of the VDR gene at the 3’ end, namely BsmI, ApaI and TaqI: the newer real-time polymerase chain reaction (qPCR) and the traditional restriction fragment length polymorphism polymerase chain reaction (RFLP-PCR) techniques. This comparison was performed to evaluate alternative methods for genotyping which consumed less time than RFLP-PCR. When the presence of each polymorphism by both the techniques was compared in this cohort of South-Asian women, it was found that RFLP-PCR proved to be a more reliable technique than qPCR for genotyping the VDR gene. Another objective of this project was to investigate the prevalence of the above three polymorphisms along with Cdx-2 and FokI SNPs which are present at the 5’ end of the VDR gene, in the population under study and their possible association with phenotypes such as vitamin D responsiveness and insulin resistance. These women were screened and biochemical data was collected during the earlier Surya Study. Of these, eighty-one women were then selected for intervention based on them having high insulin resistance (HOMA-IR>1.93) and serum 25(OH)D<50 nmol/L. Out of these eighty-one women, forty-two were given vitamin D supplement and thirty-nine were given a placebo for six months. Baseline and endpoint measurements included insulin resistance (HOMA-IR), insulin sensitivity (HOMA2%S) etc. How each individual responded to treatment in the intervention group was analysed in the context of the polymorphisms that they had. An association of insulin resistance with BsmI, ApaI and TaqI SNPs was observed in this cohort of 239 women. The response to insulin resistance in the vitamin D supplemented group significantly differed for FokI genotype compared to other genotypes. This explained why certain women responded to treatment better than the others. When the frequencies of the genotypes of these five SNPs of the VDR gene were compared to other studies of different ethnicities, the results of this study were consistent with few studies but contradictory to others. The possible reasons for these differences could be because of small sample size and different ethnicities under study due to which the frequency of alleles and hence the genotypes differed.

2’ -5’ –oligoadenylate synthetase 1 (OAS1) is an antiviral enzyme that in the presence of double-stranded RNA structures, such as viral genomes or single-stranded RNA transcripts with significant double-stranded character, converts ATP to a series of 2’ -5’ –oligoadenylates (2-5A). 2-5A promotes dimerization of latent ribonuclease (RNaseL) to form catalytically active RNaseL, a candidate hereditary prostate cancer (PCa) gene. RNaseL is anti-proliferative and promotes senescence and apoptosis in PCa cells. Genotyping analysis was completed on over 600 genomic DNA samples from African-American and Caucasian, normal and PCa subjects. Genotyping was performed to screen the following SNPs in the last exon of OAS1 (rs10774671, rs1131476, rs1051042 and rs2660) to determine splicing and linkage disequilibrium (LD) or LD decay in relation to PCa. The rs10774671 GG and AA genotypes generate isoform 1 (p46) and isoform 3 (p48), respectively and were distributed equally in the healthy population. However, in cases, the AA genotype (p46) was significantly associated with PCa risk (OR: 1.80, P-value: < 0.0001). The genotypic frequencies of rs1131476, rs1051042 and rs2660 demonstrated significant LD but showed no association to PCa risk. We also identified protective (AACA, OR =0.06612, P < 0.001) and risk (GACA, OR= 2.31, p Additionally, we utilized two genome-wide association studies analyzing OAS1 and variants found on chromosome 12 to determine their relationship with PCa susceptibility for meta-analysis: This was done to elucidate the role of OAS1 SNPs and chromosome 12 variants in a larger population cohort with PCa susceptibility for a greater understanding of gene to gene interactions. The genome wide association studies used were, the Geneva Multiethnic Genome-wide Scan of Prostate Cancer (MEC), containing 2,841 African-American samples (1,343 cases and 1,498 controls) and 1,660 Japanese/Latino samples (834 cases and 826 controls), as well as Cancer Genetic Markers of Susceptibility (CGEMS) Prostate Cancer-Primary Scan (Stage 1) - PLCO which contains 2,841 samples of European ancestry (1,172 cases and 1,157 controls). We used PLINK, a whole genome association analysis toolset, to extract data on SNPs in association with PCa.