Skip to main content
Fig. 6 | BMC Genomics

Fig. 6

From: Genome-wide enhancer annotations differ significantly in genomic distribution, evolution, and function

Fig. 6

The genomic and functional similarities between enhancer sets are not consistent. (a) Multidimensional scaling (MDS) plot of liver enhancer sets based on the Jaccard similarity of the genomic distributions (Fig. 3b). (b) MDS plot for liver enhancers based on distances calculated from molecular function (MF) Gene Ontology (GO) term semantic similarity values with GREAT (Fig. 5e). (c, d) Ranked hierarchical clustering based on the Jaccard similarities of the genomic distributions (c) of all liver enhancer sets compared to clustering based on GO semantic similarity (d). FANTOM enhancers are the most distant from all other enhancer sets in both genomic and functional similarity, but the relationships between other sets are not conserved. Red branches denote identical subtrees within the hierarchy. (e) Hierarchical clustering based on genomic Jaccard distances for all contexts and methods with annotations in each context. (f) Hierarchical clustering of all available enhancer sets based on GO term distances. Terminal branches are colored by biological context. With the exception of FANTOM enhancers, the enhancer sets’ genomic distributions are more similar within than between biological contexts. Functional similarity does not always correlate with genomic similarity, and the clustering by biological context is weaker in functional space

Back to article page