Statistical Genetics Methods for Localizing Multiple Breast Cancer Genes

摘要

The purpose of this work was to develop statistical methods for improving current statistical analysis methods of genetic linkage and linkage disequilibrium. A method to calculate support (confidence) intervals for genetic risks was extended to include uncertainties in penetrance values. For a number of variables measured on a trait, a method, principal components of heritability, was developed that combines these variables in such a way that the resulting linear combination has highest heritability. For diseases of late onset, in which often only affected sib pairs are available, a method was developed to detect pairs that are not true siblings (they may be half-sibs or be unrelated) and eliminate them from the analysis. For the analysis of linkage disequilibrium, rates of true and apparent errors were investigated analytically and by computer simulation. This showed that in extreme situations, true error rates may be four times higher than the error rates detected as mendelian inconsistencies.