Haiyan Wang

Assistant Professor

Kansas State University

NONPARAMETRIC TESTS FOR LONGITUDINAL ARRAY COMPARATIVE GENOMIC HYBRIDIZATION DATA

Kansas State University Array comparative genomic hybridization (aCGH) data are becoming commonly available for scientists to study genetic mechanisms involved in complex biological processes. Such data typically contain a large number of probes observed repeatedly over time. Due to cost concern, the number of replicates is often very limited. Effective hypothesis testing tools need to take into account of the high dimensionality and small sample sizes. In this paper, we present a set of nonparametric hypothesis testing theory to test for main and interaction effects evolved from longitudinal high dimensional aCGH arrays. The asymptotic distributions of the test statistics are obtained under the non-classical setting in which the number of independent variables is large while the number of replicates at each time point is small. The methods are based on a general model setup that allows robust inference in presence of temporal correlations for heteroscedastic high dimensional low sample size data. They provide a flexible tool for a wide range of scientists to accelerate novel gene discovery such as identification of genome regions of aberration to control tumor progression. Simulations and applications of the new methods to DNA copy number variation from Wilm's tumor relapse study will be presented.

Authors: Ke Zhang and Haiyan Wang