Abstract

First, we will report findings on systematic biases in this technology gleaned from analysis of a single chromosome from the 1000 Genomes Project Data. While it is known that coverage of whole-genome re-sequencing data is not uniform, its variation is highly correlated across multiple samples in unrelated populations with a strong platform and date effects. We will describe how some of these systematic biases affect SNP and CNV calls. Second, we will describe our first steps towards statistical solutions for these problems. Our approach is based on specific genome features that are highly correlated to differences in coverage in genome regions. Note that this work is preliminary and that the technology is moving fast. So by the time I give this talk, in 3 months, it might have nothing to do with this abstract.