ABSTRACT
Population-based sample design is the current major study design for association studies. However, many rare variants are from recent mutations in pedigrees. The inability of common variants to account for most of the supposed heritability and the low power of population-based analysis tests for the association of rare variants have led to a renewed interest in family-based design with enrichment for risk alleles to detect the association of rare variants. It is increasingly recognized that analyzing samples from populations and pedigrees separately is highly inefficient. It is natural to unify population and family study designs for association studies. This chapter focuses on the statistical methods for a unified approach to the genetic analysis of both qualitative and quantitative traits. This chapter covers (1) kinship coefficient; genome similarity matrix and heritability; (2) mixed linear models for both single quantitative trait and multiple quantitative traits with common variants; (3) mixed functional linear models for both single and multiple quantitative traits with rare variants; (4) a unified general framework for sequence-based association studies with pedigree structure and unrelated individuals, which can utilize both linkage and linkage disequilibrium information; and (5) family-based genome information content and functional principal component analysis statistics for pathway analysis.
