The epigenetic information encoded by DNA methylation is characterized by a striking balance between stability and change. This information must be transmitted from parent to daughter strand with fidelity sufficient to avoid the stochastic shifts that lead to cancer, and be malleable enough to yield differentiated cells from totipotent precursors. We have developed molecular methods to collect validated methylation patterns from both strands of individual double-stranded DNA molecules, chemical protocols that provide highly reliable information on their methylation states, and mathematical and statistical approaches to analyze these new data. The resulting double-stranded DNA methylation data contain phylogenetic snapshots of the molecular processes whereby epigenetic information is copied from parent to daughter strand, providing otherwise-unavailable insights into the fidelity and flexibility of the epigenetic information. I will discuss our new findings on the fidelity of CpG and non-CpG methylation in differentiated and undifferentiated stem cells, the in vivo properties of the DNA methyltransferases, and the implications of DNA methylation dynamics for gene-expression states within populations of cells and populations of humans.
Drug DiscoveryGenomicsInformatics