Investigation of the Histone Code using de novo Epigenetic Knockouts

The three-dimensional structure of chromatin is increasingly understood to be a major player in determining gene regulation and genome function. Interest has risen in the prospect of therapeutics that are engineered to tailor the functional role of chromatin by modulating its' 3D structure. However, while some new biochemical methods are available to precisely edit epigenetic marks in vivo, little is known about the downstream effects of these edits on chromatin physical organization. We present a physical model for genome organization that represents epigenetic marks explicitly, providing a first step towards de novo epigenetic engineering. The underlying model consists of a polymeric molecule that represents an entire chromosome, with interactions between segments described by Flory-Huggins theory and monomer identities defined according to epigenetic mark occupancies taken from ChIP-seq experiments. Following structural optimization, we demonstrate good agreement with experimentally determined chromatin organization furnished by Hi-C contact maps and FISH experiments. In the process, we determine Flory-Huggins χ interaction parameters between every pair of epigenetic marks, which encompass the net effect of all regulatory proteins that bind to epigenetic marks and accomplish chromatin folding. Through selective mutations of the χ parameters and the location of epigenetic marks, we explore the entire design space of chromatin interactions in silico and predict the modulated chromatin conformations that would result from epigenetic engineering. We quantify the H3K9me3 modification as the strongest overall contributor to compartmentalization and find that de-novo knockouts are commensurate with some well-known epigenetic mechanisms. We also provide strong support for the histone code hypothesis, suggesting strongly that chromatin structure can be predicted from epigenetic marks alone.