BMC Genomics    volume 20, Article number: 511 (2019)

DOI: 10.1186/s12864-019-5779-x

Finally, we believe that ignoring enhancer diversity impedes research progress and replication, since “what we talk about when we talk about enhancers” includes diverse sequence elements across an incompletely understood spectrum, all of which are likely important for proper gene expression. Efforts to stratify enhancers into different classes, such as poised and latent, are steps in the right direction, but are likely too coarse given our incomplete current knowledge. We suspect that a more flexible model of distal regulatory regions is appropriate, with some displaying promoter-like sequence architectures and modifications and others with distinct regulatory properties in multiple, potentially uncharacterized, dimensions. Consistent and specific definitions of the spectrum of regulatory activity and architecture are necessary for further progress in enhancer identification, successful replication, and accurate genome annotation. In the interim, we must remember that genome-wide enhancer sets generated by current approaches should be treated as what they are—incomplete snapshots of a dynamic process.