Detection of Differentially Methylated Regions Using Kernel Distance and Scan Statistics

• Activity Number: 540
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Fengjiao Hu* and Hongyan Xu and Varghese George
• Companies: Georgia Regents University and Georgia Regents University and Georgia Regents University
• Keywords: Kernel distance ; Binomial spatial scan statistic ; Differentially methylated regions ; CpG sites ; mixed-effect model

Abstract

Genomic researchers are increasingly interested in DNA methylation that is altered in disease since epigenetic changes may be susceptible to modification by environmental factors. We propose two different approaches to test for differentially methylated regions (DMRs) that account for correlations among CpG sites within DMRs, one using a kernel distance statistic and the other using a binomial spatial scan statistic. In the first approach, the kernel distance statistic is calculated as a function of the difference in methylation rates between the treatment and the control groups for each CpG site, incorporating the correlations among the sites using the kernel function. The binomial scan statistic approach compares the likelihood ratios of the two groups with moving windows along the genome, using a mixed-effect model to account for correlation among CpG sites within each window. Both methods allow for adjusting for covariates. Simulation results indicate that both methods are robust with good power and good control of Type I error. The binomial scan statistic approach appears to have higher power, while the method based on kernel distance statistic is computationally faster.