Recapitulating 3D Architecture of Chromatin using Hi-C data
Jincheol Park (Mathematical Biosciences Institute, The Ohio State University)
(April 25, 2013 10:20 AM - 11:15 AM)
Three-dimensional (3D) chromatin organizations play an important role in transcription regulation and can be used to define chromatin signatures. There is a possibility of long-range (i.e. 3D) epigenetic controls instead of the usual paradigm of promoter-cis control mechanism in distal regulation of gene. Such newly discovered mechanism may reveal novel epigenetic biomarkers for distinguishing between different cell types, for example, normal cells vs. cancer cells.
The current state-of-the-art experimental technique in 3D architecture inference is Hi-C, which was evolved from the earliest experimental protocol, Chromosome Conformation Capture (3C). In this talk, I will propose a flexible random effect model for inference from Hi-C data on constructing 3D architecture of chromatin. The model has advantages over conventional models thanks to its capability of incorporating both correlation structure and unknown sources of error into the model. Properties of the model will be presented, which will be followed by numerical simulations and applications to Hi-C human genome data.