replaced
May 17th, 2019 at 8:37pm
Note: Replaced Biorxiv
Overview
Abstract
Over the past decade, 3C-related methods, complemented by increasingly detailed microscopic views of the nucleus, have provided unprecedented insights into chromosome folding in vivo. Here, to overcome the resolution limits inherent to the majority of genome-wide chromosome architecture mapping studies, we extend a recently-developed Hi-C variant, Micro-C, to map chromosome architecture at nucleosome resolution in human embryonic stem cells and fibroblasts. Micro-C maps robustly capture well-described features of mammalian chromosome folding including A/B compartment organization, topologically associating domains (TADs), and cis interaction peaks anchored at CTCF binding sites, while also providing a detailed 1-dimensional map of nucleosome positioning and phasing genome-wide. Compared to high-resolution in situ Hi-C, Micro-C exhibits substantially improved signal-to-noise with an order of magnitude greater dynamic range, enabling not only localization of domain boundaries with single-nucleosome accuracy, but also resolving more than 20,000 additional looping interaction peaks in each cell type. Intriguingly, many of these newly-identified peaks are localized along stripe patterns and form transitive grids, consistent with their anchors being pause sites impeding the process of cohesin-dependent loop extrusion. Together, our analyses provide the highest resolution maps of chromosome folding in human cells to date, and provide a valuable resource for studies of chromosome folding mechanisms.
Authors
Nils Krietenstein • Sameer Abraham • Sergey V. Venev • Nezar Abdennur • Johan Gibcus • Tsung-Han S. Hsieh • Krishna Mohan Parsi • Liyan Yang • Ren Maehr • Leonid A. Mirny • Job Dekker • Oliver J. Rando
Link
Journal
bioRxiv
doi:10.1101/639922
Published
May 17th, 2019
Curator's Note
The Micro-C and Hi-C datasets produced in this publication comprise some of the deepest Hi-C datasets on the 4DN portal (in terms of read depth). The HFFc6 Micro-C experiment set 4DNESWST3UBH is the deepest of all the 4DN data, with 5.86 billion read pairs.
HiGlass Displays
Micro-C maps of H1 hESCs and HFFs tend to look very similar at mid-level resolutions, but at higher resolutions, close to nucleosome-level, Micro-C has a much higher signal-to-noise ratio. Below are comparisons of Micro-C and Hi-C in both cell types, first at 5kb resolution and below that at 50kb resolution.