The Human Genome Project and its spin-offs such as the Allele Frequency/Genotype Project or the HapMap Project (The International HapMap Consortium 2003) are making it increasingly feasible to disentangle the genetic basis of a given complex trait using genome-wide association studies. The goal of genetic association studies is to explain the variation in the disease trait of interest using an individual's genotype at a genetic marker. These studies take advantage of the fact that we can measure genotypes directly, either via protein electrophoretic or molecular genetic methods. Association analyses are used in many settings: looking at the effects of markers in candidate genes, fine mapping under linkage peaks and even whole genome scans.
To date, the most popular design choice for testing association between phenotype and genotype is the case-control design based on unrelated individuals. Such a design does not a priori protect against a spurious finding of association due to population admixture or stratification. Moreover, parental genotype effects or imprinting effects can only be investigated in an association analysis when parental information is included. The work of Falk and Rubinstein (1987) revolutionized genetic association tests by noting that valid tests, immune from confounding due to admixture, could be constructed using parental genotype and the genotype data from their affected offspring. Since then, several methods have been proposed for using family data (e.g., Terwilliger and Ott 1992, Spielman et al. 1993). The most commonly used method of family-based association tests is Spielman’s Transmission Disequilibrium Test (TDT; Spielman et al. 1993) in which alleles transmitted to affected offspring are compared with the expected distribution of alleles among offspring. Since its conception, many generalizations to the affected child (trio) setting with biallelic markers and the assumption of an additive genetic model have been proposed (e.g., Sham and Curti 1995, Bickeboller and Clerget- Darpoux 1995, Schaid and Li 1997, Rabinowitz and Laird 2000, Laird et al. 2000, Horvath et al. 2001).
The topic of this thesis is to compare the power of genome-wide association screening methods using non-parametric FBAT methodologies <http://www.biostat.harvard.edu/~fbat/default.html), and parametric modeling approaches, such as those available in the ABEL suite (http://www.genabel.org/).
In practice, first both viewpoints for genome-wide association screening in family-based designs need to be understood (key references in Van Steen et al 2011). Second, a simulation study is set up, similar to Aulchenko et al (2007). Third, improved GRAMMAR-like approaches (http://www.genabel.org/) are compared with improved PBAT screening approaches (Van Steen et al 2005, Ionita-Laza et al 2007, Won et al 2009). Fourth, a real-life application is performed on available genome-wide asthma data. Fifth, the results of both simulation study and real-life data application are put into a broader framework (Van Steen 2011) and the thesis is drafted.
Depending on the progress made in this project, the work may lead to a genuine scientific publication.