Single nucleotide polymorphisms (SNPs) in complex disorders: a genome-widw computational analysis

Abstract

Common complex disorders, also known as multifactorial disorders, are characterized by the interactions of multiple genetic and/or environmental factors that influence the expression of a disorder. Although such disorders often cluster in families, they do not show a clear pattern of inheritance like Mendelian disorders. With the completion of Human Genome Project in 2003, focus has shifted on the study of variability of human genome among individuals. The most common type of DNA sequence variations found in the genome, the single nucleotide polymorphisms (SNPs), are believed to play a crucial role in determining the susceptibility of an individual to complex disorders. Among them, there has been a substantial interest in the study of missense SNPs (that lead to amino acid substitution in protein), given their potential relationship with genetic disorders. Of >17 million SNPs in human genome so far known, more than one million are missense. While majority of these missense substitutions are benign, i.e. have minimal impact on the structure or function of protein; some are functional, and may lead to significant changes in protein properties. Reliable identification of functional missense SNPs may help in revealing the underlying mechanisms of genetic basis of complex disorders. Genome-wide association studies offer a potentially powerful approach to identify genetic variants that affect susceptibility to complex disorders, without making any prior assumptions about the nature of variants involved. However, the main challenge to their identification has been to carry out large studies with replication to achieve statistical significance. These studies must also take into account the potential confounding effects of hidden population substructure, and testing very large numbers of SNPs, to avoid large number of false positives. Hence, the cost-effective genome-wide analyses still requires trimming down of screening space to include only a subset of the genome. Moreover, it is anticipated that when $1000 sequencing becomes a reality, a plethora of variations will be available in the public domain, and prioritizing them will become essential to identify the variation to phenotype correlation. In this thesis, we first prioritize regions in candidate genes likely to harbor disease-associated variations, primarily focusing on protein coding exons (or the ‘exome’).