Prediction and characterizacion of enhancer clusters in an induced cell-fate transition model

Large clusters of enhancers often regulate lineage instructive genes. Whether individual enhancers of a cluster cooperate or act independently in regulating gene expression is unknown. Current algorithms that identify such clusters are limited to their input genomic dataset types and have been criticized for their specificity and sensitivity. In this thesis, we describe a new computational framework for the identification of enhancer clusters, termed SEGCOND. We applied SEGCOND to next-generation sequencing datasets derived from different time points of a B-cell to macrophage transdifferentiation model. SEGCOND predicted novel, differentiation stage-dependent enhancer clusters that display hallmarks of previously described clusters, such as association with highly expressed, lineage-specific genes. We performed single and combinatorial enhancer excisions in two predicted clusters, harboring the IRF8 and FOS genes, and monitored their effects on gene expression and transdifferentiation dynamics. This revealed that the enhancers within the two clusters either synergize or antagonize each other in a differentiation stage-dependent manner. These results highlight an unexpected functional complexity of enhancer clusters and suggest that they are more than the sum of their constituent individual parts.