3D genome conformation

Project leader:Philippe Collas
Info:The human genome is hierarchically organized in 3 dimensions, laying blueprints of developmental gene expression and cell homeostasis. Analysis of genome-wide 3D genome conformation (Hi-C) data has revealed, among other structures, topologically-associating domains (TADs), which constitute spatial units of gene regulation. The combination of Hi-C, lamin ChIP-seq data and restraint-based 3D genome modeling using our platform Chrom3D (Paulsen 2017 Genome Biol) has further led to the identification of long-range interactions between TADs forming TAD 'cliques', most of which form domains of inactive chromatin near the nuclear periphery (Paulsen 2019 Nat Genet). Other factors such as PML nuclear bodies and histone variants (e.g. histone H3.3) contribute to defining specific genomic domains (Delbarre 2013 Genome Res; Delbarre 2017 Genome Res). Our current examinations of chromatin organization in cell populations and at the single-cell level provide insights into the dynamics 3D genome conformation.

Project participants

Julia Madsen Østerbye, Aurélie Bellanger, Natalia Galigniana, Mohamed Abdelhalim, Kristin Vekterud; collaborations with Jonas Paulsen (Institute of Biosciences, University of Oslo); David Tremethick, Australian National University, Canberra; Lee H. Wong (Monash University, Melbourne); Erwan Delbarre (Oslo Metropolitan University); Jacques Grill (Institut Gustave Roussy, Paris)

Image: 3D structural model of the genome in a human fibroblast nucleus. Model generated using our Chrom3D software, from Hi-C and lamin ChIP-seq input data to infer chromosomal interactions and positioning of chromatin domains towards the nuclear periphery. Each chromosome is labeled in a different color.

Ongoing research

Interplay between genomic organizers such as the nuclear envelope, LADs and TADs on nuclear architecture in stem cells and cancer cells
Mechanism of formation and disassembly of TAD cliques
Role of histone variant H3.3 on radial chromatin organization
Computational methods for 3D genome modeling

Recent findings

Cliques of topologically-associating domains shape the 3D genome during adipogenic differentiation (Paulsen 2019 Nat Genet 51, 835-843)
Chrom3D: a computational platform for 3D genome modeling from Hi-C and lamin-genome contacts (Paulsen 2017 Genome Biol 18:21; Paulsen 2018 Nat Protoc 13, 1137-1152)
Chrom3D on github: https://github.com/Chrom3D/Chrom3D
Manifold-based optimization for 3D modeling of chromatin from sparse Hi-C data (Paulsen 2015, PLoS Comput Biol 11, e1004396)
Mechanisms of H3.3 deposition into chromatin via PML nuclear bodies (Delbarre 2013 Genome Res 23, 440-451; Ivanauskiene 2014 Genome Res 24, 1584-1594)
Identification of PML-associated domains (PADs); PML modulates H3.3 deposition for heterochromatin maintenance (Delbarre 2017 Genome Res 27, 9913-21)
Enriched Domain Detector (EDD): a domain calling algorithm to map LADs and other broad genomic domains (Lund 2014 Nucl Acids Res 42, e92)
EDD on github: https://github.com/CollasLab/edd