ISBN:  Unavailable

Category: Unavailable

Page count: 268

Abstract: Mapping and identifying genes for complex traits are difficult because there are no simple correspondences between genotypes and traits. Complex traits can be influenced by numerous factors, including genetic and environmental determinants and population or phenotypic heterogeneity. A common strategy for mapping complex traits is to conduct association studies in regions revealed in initial linkage analyses by genotyping many single nucleotide polymorphisms (SNPs). Different study designs have been used to select individuals or families in order to reduce cost, increase power, and incorporate information provided by linkage analyses to conduct more efficient association studies. In this research, we have investigated different strategies to select, on the basis of multipoint identity-by-descent (IBD) sharing and/or quantitative variables, the most informative individuals from families with multiple siblings for association studies following linkage analysis. We have made two major contributions in extending the work of Fingerlin et al. (2004). First, we evaluate all strategies in a more general setting by testing an indirect association between the phenotype (a disease phenotype or a quantitative trait) and a SNP in linkage disequilibrium with the genetic locus that controls the phenotype. Second, we have explored several strategies in the context of quantitative traits. The strategies are most useful for pre-selection of individuals for fine mapping. The same individuals can be selected across candidate genes or a region of interest. Our results demonstrate that no single selection strategy is uniformly optimal in terms of cost effectiveness and that it is critical to choose selection criteria carefully depending on the study goals, data structures, and available information. For binary traits, selecting sibs using multipoint IBD sharing is always the most powerful method for sibships containing 3 or more affected sibs. However, selection based on disease severity is the best choice when families contain only affected sib pairs. For quantitative traits, the selection based on a proposed score statistic is the best strategy across all additive models, especially when allele frequency of the causal locus is low. Other strategies should be considered for different modes of inheritance or data structure.