Modeling exome sequencing data with generalized Gaussian distribution with application to copy number variation detection

Exome sequencing provides us an effective way to discover genetic factors that might be associated with phenotypes for complex diseases. Compared with the whole-genome sequencing, exome sequencing can satisfy the high sequencing coverage requirement while under the limited budge constraint. However, due to the nature that exons are distributed sparsely along the genome, and the technical variability between samples, the analysis of exome sequencing data is complicated and direct utilization of current whole-genome sequencing targeted methods yields wrong results. In this paper, we propose a novel model to represent the exome sequencing data. Under this model, we show that the technical variability as well as random sequencing error follow the generalized Gaussian distribution. Based on this observation, we propose a method to detect the copy number variation. Studies on real data from 1000 Genomes Projects validate the proposed algorithm.