Genetic analysis using family-based populations

Abstract

Most human traits are influenced by a combination of genetic and environmental effects.

Heritability expresses the proportion of trait variance that can be explained by genetic factors, and the 1980s heralded the beginning of studies that aimed to pinpoint genetic loci that contribute to trait variation, also known as quantitative trait loci (QTLs).

Subsequently, the availability of cheap, high-resolution genotyping chips ushered in the era of genome-wide association studies (GWAS). These genetic studies have discovered many associations between single-nucleotide polymorphisms (SNPs) and complex traits, but these associations do not explain the genetic component of these traits entirely. This is known as the ‘missing heritability’ problem.

Within this thesis, 40 medically-relevant human complex traits are studied in order to identify new QTLs. These traits include eye biometric traits, blood biochemical traits and anthropometric traits measured in approximately 28,000 individuals belonging to family-based samples from the general Scottish population (the Generation Scotland study) or from population isolates from Croatian (Korčula, Vis) or Scottish (Shetland, Orkney) islands. These individuals had been genotyped using commercially-available arrays, and unobserved genotypes were imputed using the Haplotype Reference Consortium (HRC) dataset.

In parallel to standard GWAS, these traits are analysed using two other statistical genetics approaches: variance component linkage analysis and regional heritability (RH) mapping.

Each study is analysed separately, in order to detect study-specific genetic effects that may not generalise across populations. At the same time, because most traits are available in several studies, this also enables meta-analysis, which boosts the power of discovery and can reveal cross-study genetic effects.

These methods are a priori complementary to each other, exploiting different aspects of human genetic variation, such as the segregation of variants within families (identity by descent, IBD), or the presence of the same variant throughout the general population (identity by state, IBS).

The strengths and weaknesses of these methods are systematically assessed by applying them to real and simulated datasets.