DETECTING GENES OBSCURED BY HETEROGENEITY AND PLEIOTROPY

点击翻译按钮获取中文摘要

Any quantitative phenotype which has been identified as a risk factor for a common disease may reflect the genotype more closely than does the disease state. Therefore, the risk factor may provide the information needed to classify the disease into subtypes, the first stage in dissecting the genetic heterogeneity underlying a common disease. Nevertheless, the risk factors themselves are often many steps removed from the genes which affect their levels, providing multiple opportunities for the expression of genetic variation. As a consequence, the risk factors also reflect genetic heterogeneity; pleiotropy results when one gene affects more than one risk factor. The available statistical methods used to detect the genes underlying risk factor levels, or quantitative phenotypes, often produce equivocal results. Likelihood analysis usually assumes genetic homogeneity; as a consequence the likelihood method of segregation analysis frequently yields inconclusive findings and the likelihood method of linkage analysis overestimates the recombination fraction. Quantitative phenotype extensions of sib pair linkage methods lack the power of likelihood analysis. This application proposes to develop statistical methodology effective at detecting the genes underlying a set of quantitative phenotypes which reflect genetic heterogeneity and pleiotropy for three data situations: 1) only quantitative phenotypes available; 2) genotypes within a linkage may available in addition to quantitative phenotypes; 3) genotypes at candidate loci available in addition to quantitative phenotypes. New approaches combining standard statistical methodology and pedigree likelihood analysis will be developed to detect the genes. The capability of the methodology to correctly detect genes underlying quantitative phenotypes will be assessed by comparing the characteristics of genes inferred from analysis of simulated pedigree data to the characteristics of the simulated genes. The simulated data will include marker genotypes within a linkage map, candidate genotypes, and quantitative phenotypes reflecting pleiotropy and genetic heterogeneity.