Search

MotivationGenome Architecture Mapping (GAM) was recently introduced as a digestion- and ligation-free method to detect chromatin conformation. Orthogonal to existing approaches based on chromatin conformation capture (3C), GAM’s ability to capture both inter- and intra-chromosomal contacts from low amounts of input data makes it particularly well suited for allele-specific analyses in a clinical setting. Allele-specific analyses are powerful tools to investigate the effects of genetic variants on many cellular phenotypes including chromatin conformation, but require the haplotypes of the individuals under study to be known a-priori. So far however, no algorithm exists for haplotype reconstruction and phasing of genetic variants from GAM data, hindering the allele-specific analysis of chromatin contact points in non-model organisms or individuals with unknown haplotypes.ResultsWe present GAMIBHEAR, a tool for accurate haplotype reconstruction from GAM data. GAMIBHEAR aggregates allelic co-observation frequencies from GAM data and employs a GAM-specific probabilistic model of haplotype capture to optimise phasing accuracy. Using a hybrid mouse embryonic stem cell line with known haplotype structure as a benchmark dataset, we assess correctness and completeness of the reconstructed haplotypes, and demonstrate the power of GAMIBHEAR to infer accurate genome-wide haplotypes from GAM data.AvailabilityGAMIBHEAR is available as an R package under the open source GPL-2 license athttps://bitbucket.org/schwarzlab/gamibhearMaintainer:julia.markowski@mdc-berlin.de

Hi-C, split-pool recognition of interactions by tag extension (SPRITE) and genome architecture mapping (GAM) are powerful technologies utilized to probe chromatin interactions genome wide, but how faithfully they capture three-dimensional (3D) contacts and how they perform relative to each other is unclear, as no benchmark exists. Here, we compare these methods in silico in a simplified, yet controlled, framework against known 3D structures of polymer models of murine and human loci, which can recapitulate Hi-C, GAM and SPRITE experiments and multiplexed fluorescence in situ hybridization (FISH) single-molecule conformations. We find that in silico Hi-C, GAM and SPRITE bulk data are faithful to the reference 3D structures whereas single-cell data reflect strong variability among single molecules. The minimal number of cells required in replicate experiments to return statistically similar contacts is different across the technologies, being lowest in SPRITE and highest in GAM under the same conditions. Noise-to-signal levels follow an inverse power law with detection efficiency and grow with genomic distance differently among the three methods, being lowest in GAM for genomic separations >1 Mb. This Analysis reports a computational approach to implement Hi-C, SPRITE and GAM, which allows researchers to assess the performances of the three technologies to capture DNA contacts in chromatin three-dimensional models.

Neurons and oligodendrocytes are terminally differentiated cells that perform highly specialized functions, which depend on cascades of gene activation and repression to retain homeostatic control over a lifespan. Gene expression is regulated by three-dimensional (3D) genome organisation, from local levels of chromatin compaction to the organisation of topological domains and chromosome compartments. Whereas our understanding of 3D genome architecture has vastly increased in the past decade, it remains difficult to study specialized cells in their native environment without disturbing their activity. To develop the application of Genome Architecture Mapping (GAM) in small numbers of specialized cells in complex tissues, we combined GAM with immunoselection. We applied immunoGAM to map the genome architecture of specific cell populations in the juvenile/adult mouse brain: dopaminergic neurons (DNs) from the midbrain, pyramidal glutamatergic neurons (PGNs) from the hippocampus, and oligodendrocyte lineage cells (OLGs) from the cortex. We integrate 3D genome organisation with single-cell transcriptomics data, and find specific chromatin structures that relate with cell-type specific patterns of gene expression. We discover abundant changes in compartment organisation, especially a strengthening of heterochromatin compartments which establish strong contacts spanning tens of megabases, especially in brain cells. These compartments contain olfactory and taste receptor genes, which are de-repressed in a subpopulation of PGNs with molecular signatures of long-term potentiation (LTP). We also show extensive reorganisation of topological domains where activation of neuronal or oligodendrocyte genes coincides with formation of new TAD borders. Finally, we discover loss of TAD organisation, or ‘TAD melting’, at long (>1Mb) neuronal genes when they are most highly expressed. Our work shows that the 3D organisation of the genome is highly cell-type specific in terminally differentiated cells of the brain, and essential to better understand brain-specific mechanisms of gene regulation.

Chromosome conformation capture methods have identified subchromosomal structures of higher-order chromatin interactions called topologically associated domains (TADs) that are separated from each other by boundary regions1, 2. By subdividing the genome into discrete regulatory units, TADs restrict the contacts that enhancers establish with their target genes3, 4, 5. However, the mechanisms that underlie partitioning of the genome into TADs remain poorly understood. Here we show by chromosome conformation capture (capture Hi-C and 4C-seq methods) that genomic duplications in patient cells and genetically modified mice can result in the formation of new chromatin domains (neo-TADs) and that this process determines their molecular pathology. Duplications of non-coding DNA within the mouse Sox9 TAD (intra-TAD) that cause female to male sex reversal in humans6, showed increased contact of the duplicated regions within the TAD, but no change in the overall TAD structure. In contrast, overlapping duplications that extended over the next boundary into the neighbouring TAD (inter-TAD), resulted in the formation of a new chromatin domain (neo-TAD) that was isolated from the rest of the genome. As a consequence of this insulation, inter-TAD duplications had no phenotypic effect. However, incorporation of the next flanking gene, Kcnj2, in the neo-TAD resulted in ectopic contacts of Kcnj2 with the duplicated part of the Sox9 regulatory region, consecutive misexpression of Kcnj2, and a limb malformation phenotype. Our findings provide evidence that TADs are genomic regulatory units with a high degree of internal stability that can be sculptured by structural genomic variations. This process is important for the interpretation of copy number variations, as these variations are routinely detected in diagnostic tests for genetic disease and cancer. This finding also has relevance in an evolutionary setting because copy-number differences are thought to have a crucial role in the evolution of genome complexity.

Powerful technologies have been developed to probe chromatin 3D physical interactions genome-wide, such as Hi-C, GAM and SPRITE. Due to their intrinsic differences and without a benchmarking reference, it is currently difficult to assess how well each method represents the genome 3D structure and their relative performance. Here, we develop a computational approach to implement Hi-C, GAM and SPRITE in-silico to compare the three methods in a simplified, yet controlled framework against known polymer 3D structures. We test our approach on models of three 6-Mb genomic regions, around the Sox9 and the HoxD genes in mouse ES cells, and around the Epha4 gene in mouse CHLX-12 cells. The model-derived contact matrices consistently match Hi-C, GAM and SPRITE experiments. We show that in-silico Hi-C, GAM and SPRITE average data are overall faithful to the 3D structures of the polymer models. We find that the inherent variability of model single-molecule 3D conformations and experimental efficiency differently affect the contact data of the different methods. Similarly, the noise-to-signal levels vary with genomic distance differently in in-silico Hi-C, SPRITE and GAM. We benchmark the performance of each technology in bulk and in single-cell experiments, and identify the minimal number of cells required for replicates to return statistically consistent chromatin contact measures. Under the same experimental conditions, SPRITE requires the lowest number of cells, Hi-C is close to SPRITE, while GAM is the most reproducible method to capture interactions at large genomic distances.