SNP analysis to dissect human traits

Abstract

The analysis of complex human diseases has been spurred by the number of published genomic sequence variants — many identified in the course of sequencing the human genome. But, to be useful for genetic analysis, variants have to be mapped accurately, their frequencies in various populations determined, and automated high-throughput assay techniques developed. Recently proposed methods address these issues: the use of ‘reduced representation shotgun’ methods for more efficient detection of single nucleotide polymorphisms (SNPs), the employment of high-throughput genotyping techniques, the development of SNP maps that incorporate information about linkage disequilibrium, and the use of SNPs in identifying susceptibility genes for common illnesses.

Introduction

There is increasing appreciation of the role played by genetic predisposition and susceptibility in such common neurological disorders as Alzheimer's disease (AD), epilepsy, Parkinson's disease, multiple sclerosis and stroke. These disorders show familial aggregation, although the mode of inheritance is not clearly Mendelian in most cases. The role that a common genomic variant might play in susceptibility to disease is best exemplified by the role that the apolipoprotein E (APOE) ε4 allele plays in AD. The ε4 allele is highly associated with the presence of AD and with earlier age of onset of disease. It is a robust association seen in many populations studied [1]. Polymorphic variation has also been implicated in stroke and cardiovascular disease [2] and in multiple sclerosis [3]. It is increasingly clear that the risk of developing many common disorders and the metabolism of medications used to treat these conditions are substantially influenced by underlying genomic variation, although the effects of any one variant might be small.

In the 1980s, restriction enzymes were used to identify single base-pair changes in genomic DNA that result in the gain or loss of a restriction site [4]. These nucleotide variants were called ‘restriction fragment length polymorphisms’ (RFLPs) and were used in early linkage studies. In the early 1990s, RFLPs were replaced by simple tandem repeats or microsatellite markers. These markers show high levels of allelic variation, are distributed throughout the human genome, and can be efficiently amplified using PCR. Microsatellite markers have been used successfully in the positional cloning of many monogenic disease genes by linkage and allelic association [5].

In recent years, there has been a greater interest in studying the genetic basis of more common disorders, in which multiple genes of small effect are involved and for which the modes of inheritance are more complex. Standard linkage analysis using large pedigrees has only limited power to detect such small effects in these disorders [6]. Association studies using unrelated cases and controls, or using smaller family groups such as sibling pairs or ‘two parents and affected child’ trios have been proposed to be more likely to detect these small effects. Quantitative analysis and mathematical modeling have suggested that genome-wide association studies using single nucleotide polymorphisms (SNPs) are more effective than linkage analysis for identifying complex disease genes 6., 7., 8.. Such studies can then take advantage of SNPs, which are easy to type, highly abundant (found on average once per 1.3 kb in the genome) and stable (i.e. not prone to the ‘slippage’ seen with microsatellite repeats).

SNPs that are associated with disease may have a direct effect on the function of the gene in which they are located. A variant may result in an amino acid change or may alter exon–intron splicing, thereby directly modifying the relevant protein, or it may exist in a regulatory region, altering the level of expression or the stability of the mRNA. Alternatively a SNP may be in linkage disequilibrium (LD) with the ‘true’ functional variant. LD, also known as allelic association, exists when alleles at two distinct locations of the genome are more highly associated than expected. To this end, the development of SNP-based LD maps could facilitate whole-genome association studies, leading to more efficient detection of candidate susceptibility genes.

The immense interest in studies with SNPs is illustrated by a recent PubMed review of papers with the keyword ‘SNP’ from June 2000 to present: nearly 500 papers have been published. In this review, we highlight the have been published. In this review, we highlight the results from some of the most significant studies.