New regulatory terrain

Veronica Meade-Kelly, August 28th, 2013
  • This visualization, which appeared on the cover of the Aug. 22 issue
    of Nature, was rendered from real data on DNA methylation and
    depicts "regulatory terrain" of the human genome.
    Image created by Bang Wong and Michael Ziller

When readers get their hands on the print edition of their favorite scientific journal, they might not realize that the eye-catching graphic on the cover may have been conceived by some of the authors inside its pages. Many notable journals invite researchers to submit cover proposals when they notify them that their paper has been accepted for publication in an upcoming issue. Having their design selected for the cover not only adds a feather to the cap of the graphic's designers, but also lends greater visibility to the scientific work that the researchers have spent years preparing for publication.

Such was the case when Broad senior associate member Alex Meissner and Ph.D. student Michael Ziller heard that their paper, titled, "Charting a dynamic DNA methylation landscape of the human genome," was accepted for publication in Nature. When they found out, the pair approached Broad creative director Bang Wong to help them conceptualize a cover that would help convey the main points of their research, and might appeal to the editorial team at the journal. The product of their collaboration was selected by Nature to grace the cover of the August 22 issue.

Consisting of peaks and valleys of a seemingly mountainous landscape, the cover image depicts "regulatory terrain" of the human genome as seen from a methylation perspective. Meissner's lab is involved in mapping gene regulatory regions – that is, parts of the genome actively involved in turning genes on and off.

For their new Nature paper, Meissner, Ziller, and their team were looking to chart sites across the human genome that change their DNA methylation state. DNA methylation occurs when cytosine (the letter "C" in the four-letter genetic code) is modified and gains a methyl group. Depending on the context, this small chemical modification can influence the cell’s regulatory machinery and the expression of its genes. Normal methylation changes are essential for development – but when abnormal changes occur, they can lead to diseases, including cancer.

The team searched across the genome, and in a variety of cell types, for regions where DNA methylation states changed over the course of normal development. Typically, methylation occurs on cytosines that are followed by guanine in the genetic code – that is, a "C" is followed by a "G" – so Meissner's team looked at all 28 million of those C-G sites in the genome, searching for those that changed their state in one cell type but not in others.

Ziller used the data from their genome survey to render the 3-D map that formed the basis of the cover image. The team plotted maximum change of each individual C-G on the x-axis (or left edge), and the median change for each C-G on the y-axis (or right edge). He calculated the median by determining, across all cell types that they tested, whether the C-G was typically methylated or typically unmethylated. The z-axis (the height of the terrain) represents the frequency with which these C-Gs occurred.

The image reveals three quite noticeable peaks. The highest peak depicts the most frequently occurring C-Gs: static C-Gs that never change and are always highly methylated. The next most pronounced peak represents C-Gs that never change, and are always unmethylated. The third peak, which appears more as a large, rolling hill in the landscape, consists of the dynamic C-Gs that the team focused on in their paper. This is the relatively small fraction of the genome -- about one fifth -- that changes DNA methylation states.

The team's survey of the C-G landscape revealed that these dynamic sites weren't randomly distributed, but overlapped heavily with known regulatory regions. This reaffirmed that these sites were likely either involved in or correlated with regulatory elements in the genome that, when disrupted, play a role in disease. The finding suggests that this smaller subset of C-Gs would be the most informative DNA methylation sites to sample when conducting disease research.

"We hope that the genomic regions we identified in this paper provide the basis for more focused profiling approaches that will allow researchers to more effectively investigate the most relevant regions, as opposed to spending the time and money required to profile the entire genome," Ziller said.

As for the image, the team hopes that it will be informative in its own way. Like all scientific images, it is meant to add a new perspective through which an audience can grasp complex, scientific concepts. In this case, it turns the data from the study into a design that is visually familiar – a topographical map.

Wong, who helped turn Ziller's 3-D rendering into the colorful, eye-pleasing graphic seen on the cover, suspects that the image appealed to the journal because it effectively represented the data, as well as the paper’s story: the discovery of dynamic regions in newly charted terrain. He also thinks that the image was successful because it was designed with Nature in mind.

“Each journal has its own style and preferences,” he said. “I think the artwork appealed to Nature because it was rendered from real data, composed with a decent amount of visual impact, and fit their editorial style.”

Paper cited: 

Ziller, M. et al. "Charting a dynamic DNA methylation landscape of the human genome." Nature. 500, 477–481 (22 August 2013). DOI:10.1038/nature12433