Poster: Robust estimation of DNA methylation with local regression

Cytosine DNA methylation plays an essential role in regulating gene expression and silencing transposons in various animals, plants and fungi. Recent advances in next-generation DNA sequencing technologies have enabled to observe genome-wide DNA methylation status; especially, bisulfite sequencing allows us to collect DNA methylation status at single-base resolution. However, one of major problems in this analysis is low sequence coverage of a large portion of genomes due to some biases such as the difficulty in sequencing GC-rich regions, which makes existing methods difficult to measure DNA methylation state accurately. To settle this problem, we herein propose a new analytical method based on local regression so as to detect methylation clusters properly and provide a statistical ground for measuring their DNA methylation levels. The effectiveness of this approach has been confirmed using bisulfite sequencing data of Arabidopsis thaliana. Indeed, the results were consistent with previous studies. Overall, even in the presence of sequence coverage bias, this method is robust in estimating DNA methylation pattern statistically, thereby contributing the in-depth understanding of DNA methylation function.