Poster abstracts
Poster number 40 submitted by Blythe Moreland
Model of pulldown alignments from SssI-treated DNA can improve methylation quantification from MBD enrichment-capture experiments
Blythe S. Moreland (Department of Physics, The Ohio State University), Kenji M. Oman (Fred Hutchinson Cancer Research Center), Ralf A. Bundschuh (Department of Physics, Department of Internal Medicine, Division of Hematology, Department of Chemistry and Biochemistry, The Ohio State University)
Abstract:
Pulldown experiments that use methyl-binding domain (MBD) proteins to preferentially bind to methylated CpGs and enrich a sample for highly methylated DNA remain a popular tool for probing a cell’s methylome. The conversion from the number of aligned reads at a genomic window to an absolute methylation level has been improved in the past by the computational tool Baymeth, which draws additional information from pulldown experiments done on an SssI-treated control sample. We circumvent the need for this additional sample by building a model for the pulldown probability at each genomic location. Using empirical data on the dependence of fragment length and CpG number to pulldown probability, we can model an SssI-control pulldown data set well enough to improve DNA methylation predictions in the Baymeth framework beyond those that do not use an SssI-treated control sample, and in many cases beyond those that do.
Keywords: DNA Methylation, MBD pulldown