This week's Mathematical Sciences Department colloquium is given by Dr. Shuying Sun, Assistant Professor of Case Comprehensive Cancer Center and Department of Epidemiology and Biostatistics, Case Western Reserve University
DNA methylation is a common and important molecular change that plays a key regulatory role in both normal and diseased cells. It is significant to study DNA methylation, especially differential methylation patterns between two groups of samples (e.g., normal individuals vs. patients). With next generation sequencing (NGS) technologies, it is now possible to study methylation patterns by considering methylation at all CG sites in an entire genome. However, it is challenging to analyze large and complex NGS data. In order to address this problem, we have developed a new statistical approach named HMM-Fisher, which uses a hidden Markov model and the Fisher’s exact test. In particular, we first use a hidden Markov chain to model the methylation signals by inferring the methylation state as No-methylation (N), Partial-methylation (P), or Full-methylation (F) for each sample. We then use the Fisher’s exact test to identify differentially methylated CG
sites. The advantages of HMM-Fisher are that it can incorporate neighboring CG site methylation information and reduce the impact of sequencing errors. In this poster, we show our HMM-Fisher method and compare it with the two-sample T-test using a publicly available data set.
Coffee will be served in FO 2.610F at 1:30 PM.
Please visit our webpage at http://www.utdallas.edu/math/news/ for up-to-date colloquia information.