A0630
Title: SNP-set clustering and testing with Hamming distance for association studies
Authors: Chuhsing Kate Hsiao - National Taiwan University (Taiwan) [presenting]
Charlotte Wang - Tamkang University (Taiwan)
Abstract: The computational complexity due to the large number of genomic markers collected in association studies has raised challenges in statistical inference. One common approach is to reduce the number of markers via clustering techniques. These clusters are often termed as marker-set or SNP-set in genome-wise association studies. We employ the Hamming distance to measure the similarity between strings of SNP genotypes and evaluate whether the given SNPs or SNP-sets should be clustered. We next propose a test to examine if the resulting SNP-set is associated with the disease of interest. This test measures if the similarity, based on Hamming distance again, between two individuals of different disease status differs significantly from two individuals in the same disease group. The proposed methods are illustrated with applications and simulation studies. The results show that the clustering algorithm can effectively improve the power of any association test and the test outperforms other tests especially when the number of neutral SNPs is large.
