Abstract
Previous works on Parkinson's disease (PD) mainly focused on genes differentially expressed between the anterior and the posterior sections of the brains of a normal mouse and the one with PD. However, no work has been done in finding a hierarchical pattern of gene expression between the different regions of a brain. Such a hierarchy is useful to locate genetic specializations within a normal brain, thus in analyzing how brain infirmities affect these specializations. We use a recently proposed method of robust hierarchical clustering using arithmeticharmonic cut to construct the hierarchical relation between different regions of the brain. Then, we show how similar regions of the normal and PD brain differ in gene expressions, indicating a functional variation due to Parkinson's disease in a few high- level clusters of brain regions.