Computational Identification of Genomic Features That Influence 3D Chromatin Domain Formation

Machine and Deep Learning Methods for Predicting 3D Genome Organization

Three-dimensional (3D) chromatin interactions, such as enhancer-promoter interactions (EPIs), loops, topologically associating domains (TADs), and A/B compartments, play critical roles in a wide range of cellular processes by regulating gene expression. Recent development of chromatin conformation capture technologies has enabled genome-wide profiling of various 3D structures, even with single cells. However, current catalogs of 3D structures remain incomplete and unreliable due to differences in technology, tools, and low data resolution. Machine learning methods have emerged as an alternative to obtain missing 3D interactions and/or improve resolution. Such methods frequently use genome annotation data (ChIP-seq, DNAse-seq, etc.), DNA sequencing information (k-mers and transcription factor binding site (TFBS) motifs), and other genomic properties to learn the associations between genomic features and chromatin interactions. In this review, we discuss computational tools for predicting three types of 3D interactions (EPIs, chromatin interactions, and TAD boundaries) and analyze their pros and cons. We also point out obstacles to the computational prediction of 3D interactions and suggest future research directions.

A Functional Survey of the Regulatory Landscape of Estrogen Receptor-Positive Breast Cancer Evolution

Only a handful of somatic alterations have been linked to endocrine therapy resistance in hormone-dependent breast cancer, potentially explaining ∼40% of relapses. If other mechanisms underlie the evolution of hormone-dependent breast cancer under adjuvant therapy is currently unknown. In this work, we employ functional genomics to dissect the contribution of cis-regulatory elements (CRE) to cancer evolution by focusing on 12 megabases of noncoding DNA, including clonal enhancers, gene promoters, and boundaries of topologically associating domains. Parallel epigenetic perturbation (CRISPRi) in vitro reveals context-dependent roles for many of these CREs, with a specific impact on dormancy entrance and endocrine therapy resistance. Profiling of CRE somatic alterations in a unique, longitudinal cohort of patients treated with endocrine therapies identifies a limited set of noncoding changes potentially involved in therapy resistance. Overall, our data uncover how endocrine therapies trigger the emergence of transient features which could ultimately be exploited to hinder the adaptive process. Significance: This study shows that cells adapting to endocrine therapies undergo changes in the usage or regulatory regions. Dormant cells are less vulnerable to regulatory perturbation but gain transient dependencies which can be exploited to decrease the formation of dormant persisters.

Machine and deep learning methods for predicting 3D genome organization

Three-Dimensional (3D) chromatin interactions, such as enhancer-promoter interactions (EPIs), loops, Topologically Associating Domains (TADs), and A/B compartments play critical roles in a wide range of cellular processes by regulating gene expression. Recent development of chromatin conformation capture technologies has enabled genome-wide profiling of various 3D structures, even with single cells. However, current catalogs of 3D structures remain incomplete and unreliable due to differences in technology, tools, and low data resolution. Machine learning methods have emerged as an alternative to obtain missing 3D interactions and/or improve resolution. Such methods frequently use genome annotation data (ChIP-seq, DNAse-seq, etc.), DNA sequencing information (k-mers, Transcription Factor Binding Site (TFBS) motifs), and other genomic properties to learn the associations between genomic features and chromatin interactions. In this review, we discuss computational tools for predicting three types of 3D interactions (EPIs, chromatin interactions, TAD boundaries) and analyze their pros and cons. We also point out obstacles of computational prediction of 3D interactions and suggest future research directions.

Topologically associating domain boundaries are required for normal genome function

Topologically associating domain (TAD) boundaries partition the genome into distinct regulatory territories. Anecdotal evidence suggests that their disruption may interfere with normal gene expression and cause disease phenotypes^1-3, but the overall extent to which this occurs remains unknown. Here we demonstrate that targeted deletions of TAD boundaries cause a range of disruptions to normal in vivo genome function and organismal development. We used CRISPR genome editing in mice to individually delete eight TAD boundaries (11-80 kb in size) from the genome. All deletions examined resulted in detectable molecular or organismal phenotypes, which included altered chromatin interactions or gene expression, reduced viability, and anatomical phenotypes. We observed changes in local 3D chromatin architecture in 7 of 8 (88%) cases, including the merging of TADs and altered contact frequencies within TADs adjacent to the deleted boundary. For 5 of 8 (63%) loci examined, boundary deletions were associated with increased embryonic lethality or other developmental phenotypes. For example, a TAD boundary deletion near Smad3/Smad6 caused complete embryonic lethality, while a deletion near Tbx5/Lhx5 resulted in a severe lung malformation. Our findings demonstrate the importance of TAD boundary sequences for in vivo genome function and reinforce the critical need to carefully consider the potential pathogenicity of noncoding deletions affecting TAD boundaries in clinical genetics screening.

The chromatin factor ROW cooperates with BEAF-32 in regulating long-range inducible genes

Insulator proteins located at the boundaries of topological associated domains (TAD) are involved in higher-order chromatin organization and transcription regulation. However, it is still not clear how long-range contacts contribute to transcriptional regulation. Here, we show that relative-of-WOC (ROW) is essential for the long-range transcription regulation mediated by the boundary element-associated factor of 32kD (BEAF-32). We find that ROW physically interacts with heterochromatin proteins (HP1b and HP1c) and the insulator protein (BEAF-32). These proteins interact at TAD boundaries where ROW, through its AT-hook motifs, binds AT-rich sequences flanked by BEAF-32-binding sites and motifs. Knockdown of row downregulates genes that are long-range targets of BEAF-32 and bound indirectly by ROW (without binding motif). Analyses of high-throughput chromosome conformation capture (Hi-C) data reveal long-range interactions between promoters of housekeeping genes bound directly by ROW and promoters of developmental genes bound indirectly by ROW. Thus, our results show cooperation between BEAF-32 and the ROW complex, including HP1 proteins, to regulate the transcription of developmental and inducible genes through long-range interactions.

Learning representations of chromatin contacts using a recurrent neural network identifies genomic drivers of conformation

Despite the availability of chromatin conformation capture experiments, discerning the relationship between the 1D genome and 3D conformation remains a challenge, which limits our understanding of their affect on gene expression and disease. We propose Hi-C-LSTM, a method that produces low-dimensional latent representations that summarize intra-chromosomal Hi-C contacts via a recurrent long short-term memory neural network model. We find that these representations contain all the information needed to recreate the observed Hi-C matrix with high accuracy, outperforming existing methods. These representations enable the identification of a variety of conformation-defining genomic elements, including nuclear compartments and conformation-related transcription factors. They furthermore enable in-silico perturbation experiments that measure the influence of cis-regulatory elements on conformation.

TADreg: a versatile regression framework for TAD identification, differential analysis and rearranged 3D genome prediction

Background/Aim

In higher eukaryotes, the three-dimensional (3D) organization of the genome is intimately related to numerous key biological functions including gene expression, DNA repair and DNA replication regulations. Alteration of 3D organization, in particular topologically associating domains (TADs), is detrimental to the organism and can give rise to a broad range of diseases such as cancers.

Methods

Here, we propose a versatile regression framework which not only identifies TADs in a fast and accurate manner, but also detects differential TAD borders across conditions for which few methods exist, and predicts 3D genome reorganization after chromosomal rearrangement. Moreover, the framework is biologically meaningful, has an intuitive interpretation and is easy to visualize.

Result and conclusion

The novel regression ranks among top TAD callers. Moreover, it identifies new features of the genome we called TAD facilitators, and that are enriched with specific transcription factors. It also unveils the importance of cell-type specific transcription factors in establishing novel TAD borders during neuronal differentiation. Lastly, it compares favorably with the state-of-the-art method for predicting rearranged 3D genome.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-022-04614-0.

Mechanisms of CP190 Interaction with Architectural Proteins in Drosophila Melanogaster

Most of the known Drosophila architectural proteins interact with an important cofactor, CP190, that contains three domains (BTB, M, and D) that are involved in protein–protein interactions. The highly conserved N-terminal CP190 BTB domain forms a stable homodimer that interacts with unstructured regions in the three best-characterized architectural proteins: dCTCF, Su(Hw), and Pita. Here, we identified two new CP190 partners, CG4730 and CG31365, that interact with the BTB domain. The CP190 BTB resembles the previously characterized human BCL6 BTB domain, which uses its hydrophobic groove to specifically associate with unstructured regions of several transcriptional repressors. Using GST pull-down and yeast two-hybrid assays, we demonstrated that mutations in the hydrophobic groove strongly affect the affinity of CP190 BTB for the architectural proteins. In the yeast two-hybrid assay, we found that architectural proteins use various mechanisms to improve the efficiency of interaction with CP190. Pita and Su(Hw) have two unstructured regions that appear to simultaneously interact with hydrophobic grooves in the BTB dimer. In dCTCF and CG31365, two adjacent regions interact simultaneously with the hydrophobic groove of the BTB and the M domain of CP190. Finally, CG4730 interacts with the BTB, M, and D domains of CP190 simultaneously. These results suggest that architectural proteins use different mechanisms to increase the efficiency of interaction with CP190.

3D genome organisation in Drosophila

Ever since Thomas Hunt Morgan's discovery of the chromosomal basis of inheritance by using Drosophila melanogaster as a model organism, the fruit fly has remained an essential model system in studies of genome biology, including chromatin organisation. Very much as in vertebrates, in Drosophila, the genome is organised in territories, compartments and topologically associating domains (TADs). However, these domains might be formed through a slightly different mechanism than in vertebrates due to the presence of a large and potentially redundant set of insulator proteins and the minor role of dCTCF in TAD boundary formation. Here, we review the different levels of chromatin organisation in Drosophila and discuss mechanisms and factors that might be involved in TAD formation. The dynamics of TADs and enhancer-promoter interactions in the context of transcription are covered in the light of currently conflicting results. Finally, we illustrate the value of polymer modelling approaches to infer the principles governing the three-dimensional organisation of the Drosophila genome.

The emerging role of chromatin remodelers in neurodevelopmental disorders: a developmental perspective

Neurodevelopmental disorders (NDDs), including intellectual disability (ID) and autism spectrum disorders (ASD), are a large group of disorders in which early insults during brain development result in a wide and heterogeneous spectrum of clinical diagnoses. Mutations in genes coding for chromatin remodelers are overrepresented in NDD cohorts, pointing towards epigenetics as a convergent pathogenic pathway between these disorders. In this review we detail the role of NDD-associated chromatin remodelers during the developmental continuum of progenitor expansion, differentiation, cell-type specification, migration and maturation. We discuss how defects in chromatin remodelling during these early developmental time points compound over time and result in impaired brain circuit establishment. In particular, we focus on their role in the three largest cell populations: glutamatergic neurons, GABAergic neurons, and glia cells. An in-depth understanding of the spatiotemporal role of chromatin remodelers during neurodevelopment can contribute to the identification of molecular targets for treatment strategies.

A Comparative Study of Supervised Machine Learning Algorithms for the Prediction of Long-Range Chromatin Interactions

The role of three-dimensional genome organization as a critical regulator of gene expression has become increasingly clear over the last decade. Most of our understanding of this association comes from the study of long range chromatin interaction maps provided by Chromatin Conformation Capture-based techniques, which have greatly improved in recent years. Since these procedures are experimentally laborious and expensive, in silico prediction has emerged as an alternative strategy to generate virtual maps in cell types and conditions for which experimental data of chromatin interactions is not available. Several methods have been based on predictive models trained on one-dimensional (1D) sequencing features, yielding promising results. However, different approaches vary both in the way they model chromatin interactions and in the machine learning-based strategy they rely on, making it challenging to carry out performance comparison of existing methods. In this study, we use publicly available 1D sequencing signals to model cohesin-mediated chromatin interactions in two human cell lines and evaluate the prediction performance of six popular machine learning algorithms: decision trees, random forests, gradient boosting, support vector machines, multi-layer perceptron and deep learning. Our approach accurately predicts long-range interactions and reveals that gradient boosting significantly outperforms the other five methods, yielding accuracies of about 95%. We show that chromatin features in close genomic proximity to the anchors cover most of the predictive information, as has been previously reported. Moreover, we demonstrate that gradient boosting models trained with different subsets of chromatin features, unlike the other methods tested, are able to produce accurate predictions. In this regard, and besides architectural proteins, transcription factors are shown to be highly informative. Our study provides a framework for the systematic prediction of long-range chromatin interactions, identifies gradient boosting as the best suited algorithm for this task and highlights cell-type specific binding of transcription factors at the anchors as important determinants of chromatin wiring mediated by cohesin.

ZNF143 in Chromatin Looping and Gene Regulation

ZNF143, a human homolog of the transcriptional activator Staf, is a C2H2-type protein consisting of seven zinc finger domains. As a transcription factor (TF), ZNF143 is sequence specifically binding to chromatin and activates the expression of protein-coding and non-coding genes on a genome scale. Although it is ubiquitous expressed, its expression in cancer cells and tissues is usually higher than that in normal cells and tissues. Therefore, abnormal expression of ZNF143 is related to cancer cell survival, proliferation, differentiation, migration, and invasion, suggesting that new small molecules can be designed by targeting ZNF143 as it may be a good potential biomarker and therapeutic target for related cancers. However, the mechanism on how ZNF143 regulates its targeting gene remains unclear. Recently, with the development of chromatin conformation capture (3C) and its derivatives, and high-throughput sequencing technology, new findings have been obtained in the study of ZNF143. Pioneering studies have showed that ZNF143 binds directly to promoters and contributes to chromatin interactions connecting promoters to distal regulatory elements, such as enhancers. Further, it has proved that ZNF143 is involved in CCCTC-binding factor (CTCF) in establishing the conserved chromatin loops by cooperating with cohesin and other partners. These results indicate that ZNF143 is a key loop formation factor. In addition, we report ZNF143 is dynamically bound to chromatin during the cell cycle demonstrated that it is a potential mitotic bookmarking factor. It may be associated with CTCF for mitosis-to-G1 phase transition and chromatin loop re-establishment in early G1 phase. In the future, researchers could further clarify the fine mechanism of ZNF143 in mediating chromatin loops with the help of CUT&RUN (CUT&Tag) and Cut-C technology. Thus, in this review, we summarize the research progress of TF ZNF143 in detail and also predict the potential functions of ZNF143 in cell fate and identity based on our recent discoveries.

MAGIC: A tool for predicting transcription factors and cofactors driving gene sets using ENCODE data

Transcriptomic profiling is an immensely powerful hypothesis generating tool. However, accurately predicting the transcription factors (TFs) and cofactors that drive transcriptomic differences between samples is challenging. A number of algorithms draw on ChIP-seq tracks to define TFs and cofactors behind gene changes. These approaches assign TFs and cofactors to genes via a binary designation of 'target', or 'non-target' followed by Fisher Exact Tests to assess enrichment of TFs and cofactors. ENCODE archives 2314 ChIP-seq tracks of 684 TFs and cofactors assayed across a 117 human cell lines under a multitude of growth and maintenance conditions. The algorithm presented herein, Mining Algorithm for GenetIc Controllers (MAGIC), uses ENCODE ChIP-seq data to look for statistical enrichment of TFs and cofactors in gene bodies and flanking regions in gene lists without an a priori binary classification of genes as targets or non-targets. When compared to other TF mining resources, MAGIC displayed favourable performance in predicting TFs and cofactors that drive gene changes in 4 settings: 1) A cell line expressing or lacking single TF, 2) Breast tumors divided along PAM50 designations 3) Whole brain samples from WT mice or mice lacking a single TF in a particular neuronal subtype 4) Single cell RNAseq analysis of neurons divided by Immediate Early Gene expression levels. In summary, MAGIC is a standalone application that produces meaningful predictions of TFs and cofactors in transcriptomic experiments.

Two paralogous znf143 genes in zebrafish encode transcriptional activator proteins with similar functions but expressed at different levels during early development

BACKGROUND

ZNF143 is an important transcriptional regulator protein conserved in metazoans and estimated to bind over 2000 promoter regions of both messenger RNA and small nuclear RNA genes. The use of zebrafish is a useful model system to study vertebrate gene expression and development. Here we characterize znf143a, a novel paralog of znf143b, previously known simply as znf143 in zebrafish. This study reveals a comparison of quantitative and spatial expression patterns, transcriptional activity, and a knockdown analysis of both ZNF143 proteins.

RESULTS

ZNF143a and ZNF143b have a fairly strong conservation with 65% amino acid sequence identity, and both are potent activators in transient transfection experiments. In situ hybridization analyses of both znf143 mRNAs show that these genes are expressed strongly in regions of the brain at 24 h post fertilization in zebrafish development. A transient knockdown analysis of znf143 expression from either gene using CRISPR interference revealed similar morphological defects in brain development, and caused brain abnormalities in up to 50% of injected embryos. Although present in the same tissues, znf143a is expressed at a higher level in early development which might confer an evolutionary benefit for the maintenance of two paralogs in zebrafish.

CONCLUSIONS

znf143a encodes a strong activator protein with high expression in neural tissues during early embryogenesis in zebrafish. Similar to its paralogous gene, znf143b, both znf143 genes are required for normal development in zebrafish.

TAD-free analysis of architectural proteins and insulators

The three-dimensional (3D) organization of the genome is intimately related to numerous key biological functions including gene expression and DNA replication regulations. The mechanisms by which molecular drivers functionally organize the 3D genome, such as topologically associating domains (TADs), remain to be explored. Current approaches consist in assessing the enrichments or influences of proteins at TAD borders. Here, we propose a TAD-free model to directly estimate the blocking effects of architectural proteins, insulators and DNA motifs on long-range contacts, making the model intuitive and biologically meaningful. In addition, the model allows analyzing the whole Hi-C information content (2D information) instead of only focusing on TAD borders (1D information). The model outperforms multiple logistic regression at TAD borders in terms of parameter estimation accuracy and is validated by enhancer-blocking assays. In Drosophila, the results support the insulating role of simple sequence repeats and suggest that the blocking effects depend on the number of repeats. Motif analysis uncovered the roles of the transcriptional factors pannier and tramtrack in blocking long-range contacts. In human, the results suggest that the blocking effects of the well-known architectural proteins CTCF, cohesin and ZNF143 depend on the distance between loci, where each protein may participate at different scales of the 3D chromatin organization.

Three-dimensional organization and dynamics of the genome

Genome is a complex hierarchical structure, and its spatial organization plays an important role in its function. Chromatin loops and topological domains form the basic structural units of this multiscale organization and are essential to orchestrate complex regulatory networks and transcription mechanisms. They also form higher-order structures such as chromosomal compartments and chromosome territories. Each level of this intrinsic architecture is governed by principles and mechanisms that we only start to understand. In this review, we summarize the current view of the genome architecture on the scales ranging from chromatin loops to the whole genome. We describe cell-to-cell variability, links between genome reorganization and various genomic processes, such as chromosome X inactivation and cell differentiation, and the interplay between different experimental techniques.

The 3D genome: From fundamental principles to disease and cancer

In higher eukaryotes, the three-dimensional (3D) organization of the genome is intimately related to numerous key biological functions including gene expression, DNA repair and DNA replication regulations. Alteration of this 3D organization is detrimental to the organism and can give rise to a broad range of diseases such as cancers. Here, we review recent advances in the field. We first describe how the genome is packed in 3D to form chromosome territories, compartments and domains. We also give an overview of the recent techniques that allow to map the genome in 3D up to the kilobase resolution. We then discuss potential mechanisms by which genome misfolding can affect proper gene expression by distal enhancers, and how the 3D genome influences the formation of genomic rearrangements.

How epigenome drives chromatin folding and dynamics, insights from efficient coarse-grained models of chromosomes

The 3D organization of chromosomes is crucial for regulating gene expression and cell function. Many experimental and polymer modeling efforts are dedicated to deciphering the mechanistic principles behind chromosome folding. Chromosomes are long and densely packed-topologically constrained-polymers. The main challenges are therefore to develop adequate models and simulation methods to investigate properly the multi spatio-temporal scales of such macromolecules. Here, we proposed a generic strategy to develop efficient coarse-grained models for self-avoiding polymers on a lattice. Accounting accurately for the polymer entanglement length and the volumic density, we show that our simulation scheme not only captures the steady-state structural and dynamical properties of the system but also tracks the same dynamics at different coarse-graining. This strategy allows a strong power-law gain in numerical efficiency and offers a systematic way to define reliable coarse-grained null models for chromosomes and to go beyond the current limitations by studying long chromosomes during an extended time period with good statistics. We use our formalism to investigate in details the time evolution of the 3D organization of chromosome 3R (20 Mbp) in drosophila during one cell cycle (20 hours). We show that a combination of our coarse-graining strategy with a one-parameter block copolymer model integrating epigenomic-driven interactions quantitatively reproduce experimental data at the chromosome-scale and predict that chromatin motion is very dynamic during the cell cycle.

Brca2, Pds5 and Wapl differentially control cohesin chromosome association and function

The cohesin complex topologically encircles chromosomes and mediates sister chromatid cohesion to ensure accurate chromosome segregation upon cell division. Cohesin also participates in DNA repair and gene transcription. The Nipped-B-Mau2 protein complex loads cohesin onto chromosomes and the Pds5-Wapl complex removes cohesin. Pds5 is also essential for sister chromatid cohesion, indicating that it has functions beyond cohesin removal. The Brca2 DNA repair protein interacts with Pds5, but the roles of this complex beyond DNA repair are unknown. Here we show that Brca2 opposes Pds5 function in sister chromatid cohesion by assaying precocious sister chromatid separation in metaphase spreads of cultured cells depleted for these proteins. By genome-wide chromatin immunoprecipitation we find that Pds5 facilitates SA cohesin subunit association with DNA replication origins and that Brca2 inhibits SA binding, mirroring their effects on sister chromatid cohesion. Cohesin binding is maximal at replication origins and extends outward to occupy active genes and regulatory sequences. Pds5 and Wapl, but not Brca2, limit the distance that cohesin extends from origins, thereby determining which active genes, enhancers and silencers bind cohesin. Using RNA-seq we find that Brca2, Pds5 and Wapl influence the expression of most genes sensitive to Nipped-B and cohesin, largely in the same direction. These findings demonstrate that Brca2 regulates sister chromatid cohesion and gene expression in addition to its canonical role in DNA repair and expand the known functions of accessory proteins in cohesin's diverse functions.

Convergence of topological domain boundaries, insulators, and polytene interbands revealed by high-resolution mapping of chromatin contacts in the early Drosophila melanogaster embryo

High-throughput assays of three-dimensional interactions of chromosomes have shed considerable light on the structure of animal chromatin. Despite this progress, the precise physical nature of observed structures and the forces that govern their establishment remain poorly understood. Here we present high resolution Hi-C data from early Drosophila embryos. We demonstrate that boundaries between topological domains of various sizes map to DNA elements that resemble classical insulator elements: short genomic regions sensitive to DNase digestion that are strongly bound by known insulator proteins and are frequently located between divergent promoters. Further, we show a striking correspondence between these elements and the locations of mapped polytene interband regions. We believe it is likely this relationship between insulators, topological boundaries, and polytene interbands extends across the genome, and we therefore propose a model in which decompaction of boundary-insulator-interband regions drives the organization of interphase chromosomes by creating stable physical separation between adjacent domains.

How many roads lead to cohesinopathies?

Genetic mapping studies reveal that mutations in cohesion pathways are responsible for multispectrum developmental abnormalities termed cohesinopathies. These include Roberts syndrome (RBS), Cornelia de Lange Syndrome (CdLS), and Warsaw Breakage Syndrome (WABS). The cohesinopathies are characterized by overlapping phenotypes ranging from craniofacial deformities, limb defects, and mental retardation. Though these syndromes share a similar suite of phenotypes and arise due to mutations in a common cohesion pathway, the underlying mechanisms are currently believed to be distinct. Defects in mitotic failure and apoptosis i.e. trans DNA tethering events are believed to be the underlying cause of RBS, whereas the underlying cause of CdLS is largely modeled as occurring through defects in transcriptional processes i.e. cis DNA tethering events. Here, we review recent findings described primarily in zebrafish, paired with additional studies in other model systems, including human patient cells, which challenge the notion that cohesinopathies represent separate syndromes. We highlight numerous studies that illustrate the utility of zebrafish to provide novel insights into the phenotypes, genes affected and the possible mechanisms underlying cohesinopathies. We propose that transcriptional deregulation is the predominant mechanism through which cohesinopathies arise. Developmental Dynamics 246:881-888, 2017. © 2017 Wiley Periodicals, Inc.

IC-Finder: inferring robustly the hierarchical organization of chromatin folding

The spatial organization of the genome plays a crucial role in the regulation of gene expression. Recent experimental techniques like Hi-C have emphasized the segmentation of genomes into interaction compartments that constitute conserved functional domains participating in the maintenance of a proper cell identity. Here, we propose a novel method, IC-Finder, to identify interaction compartments (IC) from experimental Hi-C maps. IC-Finder is based on a hierarchical clustering approach that we adapted to account for the polymeric nature of chromatin. Based on a benchmark of realistic in silico Hi-C maps, we show that IC-Finder is one of the best methods in terms of reliability and is the most efficient numerically. IC-Finder proposes two original options: a probabilistic description of the inferred compartments and the possibility to explore the various hierarchies of chromatin organization. Applying the method to experimental data in fly and human, we show how the predicted segmentation may depend on the normalization scheme and how 3D compartmentalization is tightly associated with epigenomic information. IC-Finder provides a robust and generic ‘all-in-one’ tool to uncover the general principles of 3D chromatin folding and their influence on gene regulation. The software is available at http://membres-timc.imag.fr/Daniel.Jost/DJ-TIMC/Software.html.

MrTADFinder: A network modularity based approach to identify topologically associating domains in multiple resolutions

Genome-wide proximity ligation based assays such as Hi-C have revealed that eukaryotic genomes are organized into structural units called topologically associating domains (TADs). From a visual examination of the chromosomal contact map, however, it is clear that the organization of the domains is not simple or obvious. Instead, TADs exhibit various length scales and, in many cases, a nested arrangement. Here, by exploiting the resemblance between TADs in a chromosomal contact map and densely connected modules in a network, we formulate TAD identification as a network optimization problem and propose an algorithm, MrTADFinder, to identify TADs from intra-chromosomal contact maps. MrTADFinder is based on the network-science concept of modularity. A key component of it is deriving an appropriate background model for contacts in a random chain, by numerically solving a set of matrix equations. The background model preserves the observed coverage of each genomic bin as well as the distance dependence of the contact frequency for any pair of bins exhibited by the empirical map. Also, by introducing a tunable resolution parameter, MrTADFinder provides a self-consistent approach for identifying TADs at different length scales, hence the acronym "Mr" standing for Multiple Resolutions. We then apply MrTADFinder to various Hi-C datasets. The identified domain boundaries are marked by characteristic signatures in chromatin marks and transcription factors (TF) that are consistent with earlier work. Moreover, by calling TADs at different length scales, we observe that boundary signatures change with resolution, with different chromatin features having different characteristic length scales. Furthermore, we report an enrichment of HOT (high-occupancy target) regions near TAD boundaries and investigate the role of different TFs in determining boundaries at various resolutions. To further explore the interplay between TADs and epigenetic marks, as tumor mutational burden is known to be coupled to chromatin structure, we examine how somatic mutations are distributed across boundaries and find a clear stepwise pattern. Overall, MrTADFinder provides a novel computational framework to explore the multi-scale structures in Hi-C contact maps.

The accommodation of the roughly 2m of DNA in the nuclei of mammalian cells results in an intricate structure, in which the topologically associating domains (TADs) formed by densely interacting genomic regions emerge as a fundamental structural unit. Identification of TADs is essential for understanding the role of 3D genome organization in gene regulation. By viewing the chromosomal contact map as a network, TADs correspond to the densely connected regions in the network. Motivated by this mapping, we propose a novel method, MrTADFinder, to identify TADs based on the concept of modularity in network science. Using MrTADFinder, we identify domains at various resolutions, and further explore the interplay between domains and other chromatin features like transcription factors binding and histone modifications at different resolutions. Overall, MrTADFinder provides a new computational framework to investigate the multiple length scales that are built inside the organization of the genome.

Chromatin Architecture Emerges during Zygotic Genome Activation Independent of Transcription

Chromatin architecture is fundamental in regulating gene expression. To investigate when spatial genome organization is first established during development, we examined chromatin conformation during Drosophila embryogenesis and observed the emergence of chromatin architecture within a tight time window that coincides with the onset of transcription activation in the zygote. Prior to zygotic genome activation, the genome is mostly unstructured. Early expressed genes serve as nucleation sites for topologically associating domain (TAD) boundaries. Activation of gene expression coincides with the establishment of TADs throughout the genome and co-localization of housekeeping gene clusters, which remain stable in subsequent stages of development. However, the appearance of TAD boundaries is independent of transcription and requires the transcription factor Zelda for locus-specific TAD boundary insulation. These results offer insight into when spatial organization of the genome emerges and identify a key factor that helps trigger this architecture.

Uncovering direct and indirect molecular determinants of chromatin loops using a computational integrative approach

Chromosomal organization in 3D plays a central role in regulating cell-type specific transcriptional and DNA replication timing programs. Yet it remains unclear to what extent the resulting long-range contacts depend on specific molecular drivers. Here we propose a model that comprehensively assesses the influence on contacts of DNA-binding proteins, cis-regulatory elements and DNA consensus motifs. Using real data, we validate a large number of predictions for long-range contacts involving known architectural proteins and DNA motifs. Our model outperforms existing approaches including enrichment test, random forests and correlation, and it uncovers numerous novel long-range contacts in Drosophila and human. The model uncovers the orientation-dependent specificity for long-range contacts between CTCF motifs in Drosophila, highlighting its conserved property in 3D organization of metazoan genomes. Our model further unravels long-range contacts depending on co-factors recruited to DNA indirectly, as illustrated by the influence of cohesin in stabilizing long-range contacts between CTCF sites. It also reveals asymmetric contacts such as enhancer-promoter contacts that highlight opposite influences of the transcription factors EBF1, EGR1 or MEF2C depending on RNA Polymerase II pausing.