Inferring chromatin loop configurations via a loop extrusion factor (LEF) model based on measured LEF density

Chromosome conformation capture (Hi-C) quantifies chromatin organization via “contact maps” that represent the relative probability of two genomic loci to be proximal. Hi-C maps have led to the identification of topologically associating domains (TADs) as key elements of the mesoscale chromatin organization. In turn, the loop extrusion factor (LEF) model has emerged as the preferred mechanism underlying TAD formation. In this model, each LEF — e.g. the DNA-binding ATPase, cohesin — binds to chromatin and initiates loop extrusion, which proceeds until the LEF dissociates or is blocked by another LEF or by a “boundary element” (BE). BE locations and activities are critical for establishing TAD patterns, and, although some BEs have been identified, many questions concerning BE identity and activity remain, thus frustrating comparisons between Hi-C data and simulations. To bypass this knowledge deficit, we reasoned that it may be possible to use the cohesin ChIP-seq data as a proxy for LEF-BE interactions. Accordingly, we present modified LEF model simulation in which the position-dependent loop extrusion rate follows directly from cohesin ChIP-seq data. For several genomic regions of several organisms, we compare (1) the simulated LEF density to cohesin ChIP-seq data, and (2) simulated and experimental Hi-C maps. This approach has the potential to accurately characterize the dynamic loop configuration of any desired genomic region in any organism.