When you’re comparing epigenomes, it’s easy to miss the forest for the trees. Because the epigenetic “fruit” on each nucleotide can vary with factors like cell type, differentiation state, and the environment, it’s often tricky to spot patterns on the epigenome-wide scale.
In a recent Nature Biotechnology commentary, Aleksandar Milosavlijevic of the NIH Epigenomics Roadmap Data Analysis and Coordination Center at Baylor College of Medicine describes what we’ll need to keep from getting lost in the epigenomic landscape.
Comparative epigenomics could help us understand how different cell lineages are related or how the epigenome changes with disease. But first, researchers must overcome some big computational and engineering hurdles. For example, when searching for similarities among epigenomes, scientists might need to analyze thousands of loci scattered throughout the genome as a single set, if such loci are controlled by the same master regulator. To meet challenges like this one, Milosavlijevic proposes a “cyberinfrastructure” that would allow researchers to share and process epigenomic data through a large network, comparing the data against a human epigenome atlas. Milosavlijevic maps out the key concepts and requirements for this infrastructure. To keep from getting lost in your own neck of the epigenomic woods, check out all the details at Nature Biotechnology October 2010.