SEMINAR: Where plant and human biology meets: mapping methylomes as an epigenetic landscape

Visualizing next gen sequencing data and integration with other genome aligned datasets is a growing requirement for researchers engaged in molecular analysis of cells from all living systems. In Plant Energy Biology we have been developing software and data visualisation tools (www.annoj.org) both to support our own research and to engage with other researchers involved in large scale data analysis projects. The model plant Arabidopsis has been a testbed for many eukaryote genomic studies, including a collaboration between the Salk Institute and UWA to make the first single base resolution methylome map of a multi-cellular eukaryote (Cell 2008, 133, 523-536). This analysis revealed over 2.2 million methylcytosines in a 125 Mb genome, with over 40% in non-CG contexts and a clear connection between methylation status and the small RNA landscape.

Through collaboration with the San Diego Epigenome Center, part of the NIH RoadMap Epigenomics Program, a modified pipeline of analysis has now been used in human cells. In this collaboration over 1 billion reads from Illumina GA analysis of sodium bisulfite-modified DNA from IMR90 fibroblast cells and H1 stem cells have been aligned to the human genome, coupled to extensive mRNA and small RNA sequencing in the same samples, and the localisation of a series of histone H3 modifications. This represents the first large-scale integrated comparison of the human epigenome between differentiated cells and pluripotent stem cells (Nature Oct 14 2009 doi:10.1038/nature08514). Approximately 60 million methylcytosines have been identified in the human genome, and in H1 stem cells ~20% of these are in non-CG contexts (showcased using annoj at http://neomorph.salk.edu/human_methylome). Non-CG methylation patterns in H1 were suppressed at the interaction sites of Sox2, NanoG and Oct4, gene body non-CG methylation in H1 cells was positively correlated with gene expression, and suppressed CG methylation in H1 compared to IMR90 revealed a set of genes potentially involved in the transition from a pluripotent to a differentiated state.