A non-parametric regression method for linkage mapping of slow Beta EEG waves

Parametric linkage methods for quantitative trait loci are sensitive to violations in trait distributional assumptions, while non-parametric methods are relatively more robust. In this article, we perform a genome-wide scan for Slow Beta EEG waves via a non-parametric regression method based on kernel smoothing using data generated in the COGA project. We obtained statistically significant linkage signals on Chromosomes 1, 4, 5 and 15 which are close to regions associated with some alcohol related phenotypes. We also test for epistatic interactions between the regions which exhibit significant linkage. Evidence of epistasis was found between regions on Chromosomes 1 and 4 with those on Chromosome 15.