| Summary: | 博士 === 國立臺灣大學 === 農藝學研究所 === 87 === This study proposes a method which avoids the time-consuming Newton-Raphson method for mapping the quantitative trait locus（QTL）of binary and ordinal data by the interval mapping method. Two parametric bootstrap methods, the incidence parametric bootstrap method and the liability parametric bootstrap method, are proposed for the calculation of the critical value and the p-value of the hypothesis test. Moreover, two nonparametric bootstrap methods, the nonparametric bootstrap method and the permutation bootstrap method, are used for the calculation of the power of the hypothesis test, the standard errors of the estimators of parameters, the confidence intervals of the QTL effect and the QTL position. Finally, simulation of the case of the binary data, including generation of the binary QTL data, mapping of the QTL by the interval mapping method（i.e., estimating the QTL position）, estimation of the QTL effect and the threshold, and calculation of the above six bootstrap statistics, are performed.
|
|---|
