CoChat
Sign Up
Cited by 1,338Genomics and Chromatin DynamicsChromatin Remodeling and CancerChromosomal and Genetic Variations

Multiscale 3D Genome Rewiring during Mouse Neural Development

Open access
Bonev, Boyan, Mendelson Cohen, Netta, Szabo, Quentin, Fritsch, Lauriane, Papadopoulos, Giorgio L., Lubling, Yaniv, Xu, Xiaole, Lv, Xiaodan +3 more·2017·Cell
DOI 10.1016/j.cell.2017.09.043PMC PMC5651218
Open this paper in CoChat — free accountOpen PDF

Abstract

Chromosome conformation capture technologies have revealed important insights into genome folding. Yet, how spatial genome architecture is related to gene expression and cell fate remains unclear. We comprehensively mapped 3D chromatin organization during mouse neural differentiation in vitro and in vivo, generating the highest-resolution Hi-C maps available to date. We found that transcription is correlated with chromatin insulation and long-range interactions, but dCas9-mediated activation is insufficient for creating TAD boundaries de novo. Additionally, we discovered long-range contacts between gene bodies of exon-rich, active genes in all cell types. During neural differentiation, contacts between active TADs become less pronounced while inactive TADs interact more strongly. An extensive Polycomb network in stem cells is disrupted, while dynamic interactions between neural transcription factors appear in vivo. Finally, cell type-specific enhancer-promoter contacts are established concomitant to gene expression. This work shows that multiple factors influence the dynamics of chromatin interactions in development.

AI snapshot

Written from the paper’s full text.

TL;DR

Ultra-deep in situ Hi-C across mouse ES cells, neural progenitors, and cortical neurons in vitro and in vivo reveals multiscale 3D genome rewiring during neural differentiation. The study links transcription, CTCF, Polycomb, neural transcription factors, and enhancer-promoter contacts to dynamic chromatin architecture, while showing that transcription alone is not sufficient to create TAD boundaries de novo.

Key findings

  • Differentiation involves global compartment reorganization: interactions between active A-type domains decrease, while contacts within inactive B-type domains become stronger; compartment borders decrease with differentiation.
  • TADs decrease in number and increase in size; CTCF-less boundaries increase. Developmentally regulated boundaries include 259 ES-specific and 54 neural-specific boundaries. CRISPR-dCas9 activation at Zfp608 and Sox4 caused >10- to 20-fold gene activation but no significant insulation or contact-map change, indicating transcription is correlated with but not sufficient for de novo TAD boundary formation.
  • CTCF-mediated loops between convergent CTCF sites increase during differentiation despite decreased CTCF binding; ES-specific CTCF contacts and insulation are lost in NPCs, validated by 3D DNA FISH at the Zfp42 locus.
  • Active promoters show CTCF-independent insulation and interact with other active promoters across TAD boundaries; gene bodies of exon-rich, highly expressed genes form long-range contacts in cis and trans, correlating with exon number/splicing. Polycomb-mediated contacts are disrupted during differentiation and correlate with Ring1B rather than H3K27me3. In vivo, neural TFs Pax6, NeuroD2, and Tbr1 organize cell-type-specific contacts, and enhancer-promoter contacts are mostly cell-type specific and generally established concomitantly with gene expression.

Methods

  • In situ Hi-C was performed on FACS-purified G0/G1 mouse ES cells, NPCs, and CNs from an in vitro differentiation system, plus NPCs and CNs purified from E14.5 mouse neocortex in vivo. The study reports >40 billion paired-end reads and >17 billion uniquely mapped contacts, with >11 billion uniquely aligned contacts in vitro and ~3 billion per in vivo cell type; maximum resolution reached 750 bp.
  • Complementary assays included RNA-seq, ChIP-seq for chromatin marks and factors such as CTCF and Ring1B, 3D DNA FISH, and CRISPR-dCas9-mediated gene activation.
  • Hi-C analysis used fragment-end resolution, expected models generated with the Shaman R package, insulation scores, eigenvector-based A/B compartment assignment, aggregate Hi-C maps, and contact enrichment quantification.

Limitations

  • The paper states that the in vitro system recapitulates some but not all aspects of cortical development, such as neuronal migration and morphogen signaling; it also notes that whether the observed gene-body contacts are directly linked to co-transcriptional splicing remains to be seen, and whether TF contacts represent true homotypic interactions is unresolved.
  • The paper states that clear violations of a simple enhancer-promoter looping model exist, as some enhancer-promoter loops are pre-established.
  • Evident from the text: Hi-C is population-averaged and cannot resolve single-cell or allele-specific heterogeneity; FISH validates only selected loci; dCas9 activation was tested at two loci; and in vitro versus in vivo differences may limit generalization.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.cell.2017.09.043

Cites

52 of 52 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 37 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    dCas9-mediated activation is insufficient for creating TAD boundaries de novo.

    “We found that transcription is correlated with chromatin insulation and long-range interactions, but dCas9-mediated activation is insufficient for creating TAD boundaries de novo.”
    • Also stated in: § Transcription Is Correlated with but Not Sufficient to Induce Insulation at TAD Boundaries De Novo · hedged
      “Therefore, transcriptional activation per se appears insufficient to cause chromatin insulation and create a TAD boundary.”
    Quote in the paperSubstantial claim
  • Findingown data

    An extensive Polycomb network in stem cells is disrupted during neural differentiation.

    “An extensive Polycomb network in stem cells is disrupted, while dynamic interactions between neural transcription factors appear in vivo.”
    • Also stated in: § Polycomb-Mediated Interactions Are Disrupted Independently of H3K27me3 during Neural Differentiation · boosted
      “we observed a strong decrease in contact intensity between common H3K27me3 sites, which was especially notable at long distances (Figure S6A).”
    Quote in the paperSubstantial claim
  • Findingown data

    Cell type-specific enhancer-promoter contacts are established concomitant to gene expression.

    “Finally, cell type-specific enhancer-promoter contacts are established concomitant to gene expression.”
    Quote in the paperSubstantial claim
  • Findingown data

    Long-range contacts occur between gene bodies of exon-rich, active genes in all cell types.

    “Additionally, we discovered long-range contacts between gene bodies of exon-rich, active genes in all cell types.”
    • Also stated in: § Active Gene Promoters and the Bodies of Exon-Rich Genes Interact at Multiple Genomic Scales · plain
      “Finally, actively transcribed genes with a large number of exons engage in long-range chromatin loops in clusters both in cis- and in trans.”
    Quote in the paperSubstantial claim
  • Findingown data

    Transcription is correlated with chromatin insulation and long-range interactions.

    “We found that transcription is correlated with chromatin insulation and long-range interactions, but dCas9-mediated activation is insufficient for creating TAD boundaries de novo.”
    Quote in the paperSubstantial claim
  • May combine two claims. 53
  • Classified as probably not a substantial claim. 14
  • The quote only partly states this claim. 1

  • May combine two claims.

    Findingown data

    During neural differentiation, contacts between active TADs become less pronounced while inactive TADs interact more strongly.

    “During neural differentiation, contacts between active TADs become less pronounced while inactive TADs interact more strongly.”
    • Also stated in: § Global Reorganization of the 3D Genome during Neural Differentiation · plain
      “Surprisingly, the interaction strength between A-type domains decreased, whereas contacts within the B compartment became stronger between NPCs and CNs (Figures 1E and S1L).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The dataset comprises over 40 billion paired-end reads and more than 17 billion uniquely mapped contacts.

    “In total, we sequenced over 40 billion paired-end reads that resulted in more than 17 billion uniquely mapped contacts.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For Sox4, CTCF was bound to the promoter already in ESs, yet insulation dramatically increased only when the gene became expressed at the NPC stage.

    “CTCF was bound to the Sox4 promoter already in ESs, yet insulation dramatically increased only when the gene became expressed at the NPC stage (Figures S3K and S3J).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mechanisms contributing to developmentally regulated TAD borders remain poorly understood.

  • May combine two claims.

    Findingown data

    TAD number decreased during differentiation, with a corresponding increase in size.

    “We then examined TADs and found that their number decreased during differentiation (Figure S3A), with a corresponding increase in size (Figure S3B).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The correlation between signal enrichment and eigenvector value for active histone marks decreased during differentiation, while the anti-correlation between H3K9me3 and the eigenvector value became stronger.

    “the correlation between signal enrichment and eigenvector value for active histone marks decreased during differentiation, while the anti-correlation between H3K9me3 and the eigenvector value became stronger”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    A novel TAD boundary was formed at the Tss of Zfp608 during differentiation, concomitant with activation of the gene but in the absence of CTCF binding.

    “a novel TAD boundary was formed at the transcriptional start site (Tss) of Zfp608 during differentiation (Figure 2H), concomitant with the activation of the gene (Figure S3J) but in the absence of CTCF binding.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mechanisms contributing to developmentally regulated TAD borders remain poorly understood.

  • May combine two claims.

    Findingown data

    TAD boundaries fall into three classes: CTCF-positive boundaries, CTCF-negative boundaries close to active gene promoters, and boundaries without enrichment of active marks.

    “We observed three classes of boundaries: CTCF-positive boundaries, CTCF-negative boundaries close to active gene promoters, and a set of boundaries without enrichment of active marks (Figure 2A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cohesin is found both at CTCF-positive boundaries, where it colocalizes with CTCF, and on CTCF-negative borders marked by active transcription.

    “cohesin was found both at CTCF-positive boundaries, where it colocalized with CTCF, and on CTCF-negative borders marked by active transcription (Figure 2A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Absolute levels of Ring1B decreased during differentiation, with only a mild effect on gene expression, while H3K27me3 levels did not decrease.

    “the absolute levels of Ring1B decreased during differentiation (with only a mild effect on gene expression [Figure S6D]), while H3K27me3 levels did not (Figure 5D).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Both the contact enrichment and the insulation between ES-specific CTCF regions were decreased in NPCs.

    “In contrast, both the contact enrichment and the insulation between ES-specific CTCF regions were decreased in NPCs (Figures 3E, 3F and S4F).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Strong enrichment occurs between convergent CTCF sites, and loops are constrained primarily inside TADs and rarely cross TAD boundaries.

    “We observed strong enrichment as expected (Guo et al., 2015, Rao et al., 2014), and the loops were constrained primarily inside TADs and rarely crossed TAD boundaries (Figures 3B and 3D).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Guo et al., 2015 source

      Not checked in the free run.

    • cites Rao et al., 2014 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Conserved CTCF sites were characterized by strong interactions in the convergent orientation and by strong insulation both in ESs and NPCs.

    “Conserved CTCF were characterized by strong interactions in the convergent orientation (Figures 3E and 3F) and by strong insulation both in ESs and NPCs (Figure S4E).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Clusters of very long-range (>30 Mb) chromatin contacts occur between bodies of expressed genes in cis and also in trans.

    “we also identified clusters of very long-range (> 30 Mb) chromatin contacts between bodies of expressed genes in cis (Figures 4I and S5K) and also in trans (Figure S5L).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Most of these promoter interactions were not mediated by CTCF and were readily identified in all three cell types.

    “most of these interactions were not mediated by CTCF (Figure 4F) and were readily identified in all three cell types (Figure 4G).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    FISH confirmed that only the 3D distance between the pair of probes containing the forward and reverse CTCF site increased during differentiation, while the other probe pairs displayed the opposite effect.

    “only the 3D FISH distance between the pair of probes containing the forward and the reverse CTCF site increased during differentiation, while the other probe pairs displayed the opposite effect (Figures 3H, S4H, and S4I).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Contact enrichment between gene bodies is highly correlated not only with expression levels, but also with the number of exons per gene, and is evident across a large range of genomic distances in all three cell types.

    “the contact enrichment between gene bodies was highly correlated not only with expression levels, but also with the number of exons per gene (Figures 4J, 4K, and 4M), and was evident across a large range of genomic distances (Figure S5M) in all three cell types (Figure 4N).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The overall enrichment of CTCF-mediated contacts increased with neural differentiation, and even inter-domain contacts became apparent.

    “However, the overall enrichment of CTCF-mediated contacts increased with neural differentiation, and even inter-domain contacts became apparent (Figures 3B–3D and S4B).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    During differentiation, only the forward CTCF site became inaccessible to CTCF, and the chromatin loop disappeared.

    “During differentiation, only the forward CTCF site became inaccessible to CTCF, and the chromatin loop disappeared (Figure 3G).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    General TFs such as Taf3 appear to enhance interaction between promoters but had little effect on insulation.

    “general TFs such as Taf3 (Liu et al., 2011) appear to enhance interaction between promoters but had little effect on insulation (Figure S5D).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Liu et al., 2011 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    There is a strong preference for insulation centered on active Tss, but not on inactive ones, and this phenomenon is not dependent on CTCF binding.

    “we identified a strong preference for insulation centered on active Tss, but not on inactive ones, a phenomenon that was not dependent on CTCF binding (Figures 4A, 4B, S5A, and S5B).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Average intra-TAD contacts become stronger, while inter-TAD contacts appear to be depleted during differentiation.

    “We observed that average intra-TAD contacts become stronger, while inter-TAD contacts appear to be depleted during differentiation (Figures 3A and S4A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No central enrichment was observed when using 3′ gene ends, showing that these interactions were specific to promoters and correlated with expression level.

    “we observed no central enrichment when we performed the same analysis using instead 3′ gene ends (Figure S5F), showing that these interactions were specific to promoters and were correlated with the level of expression (Figures 4H and S5J).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Active chromatin loops span across domains and can be observed even at long distances, in contrast with CTCF-mediated contacts.

    “active chromatin loops span across domains and can be observed even at long distances (Figure 4E), in contrast with CTCF-mediated contacts.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Most TAD boundaries are conserved.

    “Therefore, although we confirmed that most TAD boundaries are conserved (Dixon et al., 2012), novel borders can occur at promoters of developmentally regulated genes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dixon et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Regions switch compartments between the in vitro and in vivo systems, accompanied by corresponding changes in gene expression.

    “Moreover, we identified regions that switch compartments between the in vitro and the in vivo systems, which were accompanied with corresponding changes in gene expression (Figures S8B and S8C).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The interaction between the Sox2 promoter and the ES-specific enhancer strongly decreases during differentiation, coinciding with a decrease in both CTCF binding and the H3K27ac mark on the enhancer.

    “We observed a strong decrease in the interactions between Sox2 promoter and the ES-specific enhancer (Figure 7G, blue arrow) during differentiation, which coincides with a decrease in both CTCF binding and the H3K27ac mark on the enhancer (Figure 7Gi).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    NeuroD2 and Tbr1 interactions increase during differentiation and are strongest in CNs in vivo.

    “NeuroD2 and Tbr1 interactions were also mostly constrained inside TADs (Figures 6C, 6D, S7P, and S7Q) but increased during differentiation and were the strongest in CNs in vivo.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    These interactions are strongly correlated with the number of splicing events per gene, with the strongest contacts occurring between genes that undergo most splicing.

    “These interactions are strongly correlated with the number of splicing events per gene with the strongest contacts occurring between genes that undergo most splicing.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cell differentiation is accompanied by a global chromatin reorganization and a change in compartmentalization.

    “The analysis of a well-controlled differentiation system and its in vivo cell counterparts shows that cell differentiation is accompanied by a global chromatin reorganization and a change in compartmentalization with a decrease in long-range interactions between A-type domains concomitant with an increase in contacts between inactive B-type domains.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    In ESs, strong long-range interactions involve Polycomb-bound genes and, to a lesser extent, genes bound by pluripotency transcription factors.

    “In ESs, the strong long-range interactions also involve Polycomb-bound genes and, to a lesser extent, genes bound by pluripotency TFs.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    These networks are reorganized in neuronal differentiation, which in the case of Polycomb appears to correlate with PRC1 binding levels.

    “These networks are reorganized in neuronal differentiation, which in the case of Polycomb appears to correlate with PRC1 binding levels.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    An NPC-specific enhancer interacted strongly with the promoter of the neuronal transcription factor Brn2 gene only in NPCs, and this interaction was disrupted in CNs when the enhancer was inactive.

    “We identified a NPC-specific enhancer, which interacted strongly with the promoter of the neuronal TF Brn2 gene only in NPCs, and this interaction was disrupted in CNs when the enhancer was inactive (Figure 7F).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    NPC populations, either in vitro or in vivo, are highly similar to each other and are distinct from CNs and vice versa.

    “Principal component analysis (PCA) also confirmed that NPC populations, either in vitro or in vivo, are highly similar to each other and are distinct from CNs and vice versa (Figure S8A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Shuffling the position of the enhancer region within the same TAD strongly decreases average Hi-C scores and the correlation between expression and chromatin contacts.

    “Furthermore, the average Hi-C scores and the correlation between expression and chromatin contacts were strongly decreased when we randomly shuffled the position of the enhancer region within the same TAD (Figures 7D, 7E, and S8H).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Contact enrichment between Pax6-bound sites is cell-type specific and strongest in ncx NPC, where Pax6 is highly expressed.

    “We observed a cell-type specific contact enrichment between Pax6-bound sites strongest in ncx NPC where Pax6 is highly expressed (Figures 6C and 6D).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The majority of enhancer-promoter interactions appear to be established concomitantly with gene expression and are disrupted when the genes become repressed.

    “Although we did identify instances of pre-set chromatin interactions between enhancers and their cognate target genes that precede transcriptional activation (Figures 7C, S8N, and S8O), the majority of those interactions appear to be established concomitantly to gene expression and are also disrupted when the genes become repressed.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The mean nuclear radius was 1.73 ± 0.28 μm for ES cells, 1.39 ± 0.24 μm for NPCs, and 1.43 ± 0.20 μm for CNs.

    “mean nuclear radius of the ES, NPC and CNs population were calculated (ES = 1.73 ± 0.28; NPC = 1.39 ± 0.24; CN = 1.43 ± 0.20).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    TADs in the two k-means clusters exhibit distinct signatures based on enrichment of chromatin marks, replication timing, and Lamin B1 DamID enrichment, leading to assignment of cluster 1 to the B (inactive) compartment and cluster 2 to the A (active) compartment.

    “Domains in each cluster exhibit distinct signature based on the enrichment of chromatin marks, replication timing and Lamin B1 DamID enrichment where available (Figure S1I), prompting us to assign cluster 1 to the B-, inactive compartment and cluster 2 to the A-, active compartment.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Topologically associating domains (TADs) exist at intermediate scales (100 kb to 1 Mb).

    “At intermediate scales (100 kb to 1 Mb), a major discovery is the existence of topologically associating domains (TADs) (Dixon et al., 2012, Nora et al., 2012, Sexton et al., 2012).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Dixon et al., 2012 source

      Not checked in the free run.

    • cites Nora et al., 2012 source

      Not checked in the free run.

    • cites Sexton et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    The authors comprehensively mapped 3D chromatin organization during mouse neural differentiation in vitro and in vivo, generating the highest-resolution Hi-C maps available to date.

    “We comprehensively mapped 3D chromatin organization during mouse neural differentiation in vitro and in vivo, generating the highest-resolution Hi-C maps available to date.”
    • Also stated in: § Global Reorganization of the 3D Genome during Neural Differentiation · plain
      “We reached a maximum resolution of 750 bp—the highest to date (Figure S1F).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Our understanding of the interaction dynamics at the smallest scales during developmental processes is currently limited.
    • Rests on an assumption: The relation between loop extrusion models and transcription remains unclear.

  • May combine two claims.

    Interpretationown data

    Cellular differentiation involves changes at all scales of folding, including CTCF-dependent processes and major changes in Polycomb and TF interaction networks.

    “Our data demonstrate that cellular differentiation involves changes at all scales of folding, and in addition to CTCF-dependent processes, it includes major changes in Polycomb and in transcription factor (TF) interaction networks.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Our understanding of the interaction dynamics at the smallest scales during developmental processes is currently limited.
    • Rests on an assumption: The relation between loop extrusion models and transcription remains unclear.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Interactions between domains give rise to two main types of spatial compartments referred to as A and B.

    “At larger scales, interactions between domains are thought to give rise to two main types of spatial compartments referred to as “A” and “B” (Lieberman-Aiden et al., 2009)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lieberman-Aiden et al., 2009 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    These borders represent a genuine loss or gain of insulation, respectively.

    “verified that these borders represent a genuine loss or gain of insulation, respectively, using aggregate Hi-C maps (Figure 2E), insulation score (Figures S3G and S3H), and directionality index (DI) (Figure S3I).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The mechanisms contributing to developmentally regulated TAD borders remain poorly understood.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Domains and compartment formation can be at least partially uncoupled and may result from different biological phenomena.

    “recent evidence indicates that domains and compartment formation can be at least partially uncoupled and may result from different biological phenomena (Nora et al., 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Nora et al., 2017 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Domain insulation is important for gene expression and physiology.

    “Domain insulation has been shown to be important for gene expression and physiology (Flavahan et al., 2016, Lupiáñez et al., 2015).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Flavahan et al., 2016 source

      Not checked in the free run.

    • cites Lupiáñez et al., 2015 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    A/B compartments are functionally related to gene expression and can switch extensively during differentiation.

    “which are functionally related to gene expression and can switch extensively during differentiation (Dixon et al., 2015).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dixon et al., 2015 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Very few studies using mostly in vitro cultured cells have achieved high resolution in Hi-C.

    “very few studies using mostly in vitro cultured cells have achieved high resolution (Rao et al., 2014).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Rao et al., 2014 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    FACS allowed purification of G0G1 cells based on DNA content.

    “we employed a fluorescence-activated cell sorting (FACS) approach, which allowed us to purify G0G1 cells based on DNA content”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The authors controlled for cell type heterogeneity and cell-cycle variation in their 3D chromatin mapping.

    “we controlled for cell type heterogeneity and cell-cycle variation.”
    • Also stated in: § Global Reorganization of the 3D Genome during Neural Differentiation · plain
      “we employed a fluorescence-activated cell sorting (FACS) approach, which allowed us to purify G0G1 cells based on DNA content”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Architectural proteins such as CTCF and cohesin play a role in domain formation.

    “It is known that architectural proteins such as CTCF and cohesin play a role (Haarhuis et al., 2017, Nora et al., 2017, Phillips-Cremins et al., 2013)”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Haarhuis et al., 2017 source

      Not checked in the free run.

    • cites Nora et al., 2017 source

      Not checked in the free run.

    • cites Phillips-Cremins et al., 2013 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    Biological replicates were highly correlated at all resolutions and for each cell type.

    “Biological replicates were highly correlated at all resolutions (Figure S1D) and for each different cell type (Figures S1E and S2A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    A major genome-folding reprogramming occurs during neural differentiation with a progressive increase in compartment size, decreased interactions within the A compartment, and a concomitant increase in interactions between B-type domains.

    “In conclusion, a major genome-folding reprogramming occurs during neural differentiation with a progressive increase in compartment size, decreased interactions within the A compartment, and a concomitant increase in interactions between B-type domains.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Our understanding of the interaction dynamics at the smallest scales during developmental processes is currently limited.
    • Rests on an assumption: The relation between loop extrusion models and transcription remains unclear.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Deposition of chromatin marks is thought to contribute to compartment switching.

    “The deposition of chromatin marks is thought to contribute to this process (Dixon et al., 2015, Nora et al., 2012)”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Dixon et al., 2015 source

      Not checked in the free run.

    • cites Nora et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Promoters of active genes are characterized by strong insulation, which is correlated with transcriptional activity, but is mostly CTCF independent.

    “In summary, promoters of active genes are characterized by strong insulation, which is correlated with transcriptional activity, but is mostly CTCF independent.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Polycomb-mediated interactions have been previously reported in mouse ESs using a Capture-C approach and are a prominent feature in Drosophila.

    “Polycomb-mediated interactions have been previously reported in mouse ESs using a Capture-C approach (Schoenfelder et al., 2015b) and are a prominent feature in Drosophila (Bantignies et al., 2011).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Schoenfelder et al., 2015b

      Not checked in the free run.

    • cites Bantignies et al., 2011 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Active Tss maintain contacts with the gene body and can interact with other transcribed Tss across TAD boundaries.

    “Active Tss maintain contacts with the gene body and, contrary to CTCF interactions, can interact with other transcribed Tss across TAD boundaries.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    As many as 40-50% of CTCF sites may change in different cell types.

    “As many as 40–50% of the CTCF sites may change in different cell types (Chen et al., 2012).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Chen et al., 2012 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Chromatin insulation occurs at multiple levels and leads to a hierarchical 3D nuclear organization, whereby smaller, less insulated regions (sub-TADs/contact domains) are nested inside TADs.

    “Chromatin insulation occurs at multiple levels and leads to a hierarchical 3D nuclear organization, whereby smaller, less insulated regions (called also sub-TADs [Flavahan et al., 2016, Phillips-Cremins et al., 2013] or contact domains [Rao et al., 2014]) are nested inside TADs, which are usually bigger and more strongly insulated (Dixon et al., 2012, Nora et al., 2012, Sexton et al., 2012).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Flavahan et al., 2016 source

      Not checked in the free run.

    • cites Phillips-Cremins et al., 2013 source

      Not checked in the free run.

    • cites Rao et al., 2014 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Few public ChIP-seq profiles exist for neural-specific transcription factors, and most of those have been performed in the developing mouse neocortex.

    “However, very few public chromatin immunoprecipitation sequencing (ChIP-seq) profiles exist, and most of those have been performed in the developing mouse neocortex (Bayam et al., 2015, Notwell et al., 2016, Sun et al., 2015).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Bayam et al., 2015 source

      Not checked in the free run.

    • cites Notwell et al., 2016 source

      Not checked in the free run.

    • cites Sun et al., 2015 source

      Not checked in the free run.

  • The quote only partly states this claim.

    Study design and conductown data

    The purified NPC and CN populations were verified using known markers Nestin and Dcx.

    “We performed RNA-seq and verified our approach using known markers for neural progenitors (Nestin) and differentiating neurons (Dcx) (Figure S7D).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The Hi-C data are high quality and samples group together based on cell type.

    “Pearson correlation coefficient (Figure S7E), comparison of the eigenvector values, insulation score, and contact probability (Figures S7G and S8A) confirmed the high quality of the data and that, importantly, samples group together based on cell type.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Pax6 is one of the few master regulator neural transcription factors with a well-known role in cortical development.

    “Pax6 is one of the few “master regulator” neural TFs with a well-known role in cortical development (Manuel et al., 2015).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Manuel et al., 2015 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Enhancer-promoter interactions are highly dynamic and cell-type specific.

    “These results suggest that E-P interactions are highly dynamic and cell-type specific.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Enhancer-promoter interactions are highly dynamic and are usually established concomitantly with gene expression.

    “Finally, E-P interactions are highly dynamic and are usually established concomitantly with gene expression.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    TAD boundaries were retained only if they were detected in at least three biological replicates.

    “Boundaries within 10Kb of each other were merged and only boundaries, which were detected in at least 3 biological replicates, were retained.”
    Quote in the paperSubstantial claim
65 of 65

  • The paper relies on this without support.

    Our understanding of the interaction dynamics at the smallest scales during developmental processes is currently limited.

    “However, our understanding of the interaction dynamics at the smallest scales during developmental processes is currently limited.”

  • The paper relies on this without support.

    The relation between loop extrusion models and transcription remains unclear.

    “However, the relation between such models and transcription remains unclear.”

  • The paper relies on this without support.

    The mechanisms contributing to developmentally regulated TAD borders remain poorly understood.

    “yet the mechanisms contributing to developmentally regulated borders remain poorly understood.”

  • The paper relies on this without support.

    The majority of strong Polycomb-bound regions are on gene promoters.

    “As the majority of strong Polycomb-bound regions are on gene promoters, we examined whether their interaction might depend on paused Pol II that is located at bivalent gene promoters (Min et al., 2011).”

  • The paper relies on this without support.

    Enhancer-promoter contacts have been difficult to observe in Hi-C, and their strength relative to surrounding regions is unclear.

    “Enhancer-promoter (E-P) contacts have been difficult to observe in Hi-C, and it is unclear how strong they are compared to the surrounding regions.”

  • The paper relies on this without support.

    The in vitro differentiation system recapitulates some aspects of cortical development but not others such as neuronal migration and morphogen signaling.

    “Furthermore, although the in vitro system we employed recapitulates some aspects of cortical development, other features, such as neuronal migration and morphogen signaling, are not.”

  • The paper relies on this without support.

    The vast majority of ultra-close cis Hi-C contacts (separated by less than 1 kb) originate from non-digested DNA.

    “Contacts separated by less than 1kb were not considered for subsequent analysis, as the vast majority of this ultra-close cis contacts originate from non-digested DNA.”
7 of 7

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The Sox2 gene has an essential enhancer, the Sox2 Control Region (SCR), which was found by 3C-derived techniques to loop to the Sox2 promoter.

    “The Sox2 gene has an essential enhancer, the Sox2 Control Region (SCR), which was found by 3C-derived techniques to loop to the Sox2 promoter (for example, [379, 401]).”

    From Transcriptional enhancers and their communication with gene promoters (2021) · cites this paper as [401]

    The Results section (Figure 7G; page not provided) reports Hi-C contacts between the Sox2 promoter and an embryonic-stem-cell-specific enhancer. It does not identify that enhancer as the Sox2 Control Region (SCR) or establish in this paper that it is essential. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Active promoters and enhancers can likely give rise to TAD boundaries even in the absence of convergent CTCF sites.

    “active promoters and enhancers can likely give rise to TAD boundaries even in the absence of convergent CTCF sites [401].”

    From Transcriptional enhancers and their communication with gene promoters (2021) · cites this paper as [401]

    In Results, “Transcription Is Correlated with but Not Sufficient to Induce Insulation at TAD Boundaries De Novo” (page not provided), the paper identifies CTCF-negative TAD boundaries near active promoters. It does not establish that active enhancers give rise to boundaries, and its gene-activation experiment found that transcriptional activation alone was insufficient to create one. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Hi-C data from differentiated mouse neurons detected chromatin interactions between HAR202 and the Npas3 promoter.

    “Interestingly, Hi-C data obtained using differentiated mouse neurons led us to detect chromatin interactions between HAR202 and the Npas3 promoter (Bonev et al. 2017; supplementary fig. S8, Supplementary Material online).”

    From The Human Accelerated Region HAR202 Controls <i>NPAS3</i> Expression in the Developing Forebrain Displaying Differential Enhancer Activity Between Modern and Archaic Human Sequences (2024) · cites this paper as Bonev et al. 2017

  • The majority of specific enhancer-promoter loops are highly enriched within, compared with across, TAD boundaries.

    “The majority of specific enhancer–promoter loops follow this pattern and are highly enriched within, compared with across, TAD boundaries [277, 401, 405].”

    From Transcriptional enhancers and their communication with gene promoters (2021) · cites this paper as [401]

  • The 3D genome plays an important role in neuronal gene expression and function.

    “The 3D genome plays an important role in neuronal gene expression and function”

    From Neuronal DNA double-strand breaks lead to genome structural variations and 3D genome disruption in neurodegeneration (2023) · cites this paper as [96]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Guide RNA is labeled for a different gene than the reported activation target

    Minor

    The experiment is described as activating Zfp608 and Sox4, but the methods label one of the guide RNAs "Zfp604." As written, the stated guide target does not match the gene used for the reported Zfp608 activation result.

  • Deposited-data table labels the CN RNA sample as NPC

    Minor

    The table lists both complementary sorted populations as ncx_NPC, whereas the paper identifies Hes5-GFP−/Dcx-mRFP+ cells as ncx CN.

  • In vivo sorting methods list the NPC population twice

    Minor

    The figure identifies GFP+/RFP− cells as NPCs and GFP−/RFP+ cells as CNs. The isolation methods instead list GFP+/RFP− and RFP−/GFP+-the same population in reversed notation-rather than the CN population.

  • Cacna2d1 contact distance exceeds the stated map region

    Minor

    The text places the interacting Ring1B site approximately 500 kb downstream of the Cacna2d1 promoter, but the Figure S6J caption says its map shows an approximately 150-kb region around that promoter containing the interaction. A 150-kb region cannot contain both sites as described.

4 of 4

Cites

Flagged first, then by citations

1 of the 53 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Related

Authors
Bonev, B., Mendelson Cohen, N., Szabo, Q., Fritsch, L., Papadopoulos, G. L., Lubling, Y., Xu, X., Lv, X., Hugnot, J.-P., Tanay, A., & Cavalli, G.
Affiliations
  • Centre National de la Recherche Scientifique (FR)
  • Weizmann Institute of Science (IL)
  • BGI Group (China) (CN)
  • Inserm (FR)
Corresponding author
  • Bonev, Boyan
  • Cavalli, Giacomo
ORCID
Venue
Cell
Publisher
Elsevier BV
Published
2017-10
Volume
171
Issue
3
Pages
557-572.e24
ISSN
0092-8674 (ISSN-L)
DOI
10.1016/j.cell.2017.09.043
PMC
PMC5651218
OpenAlex
W2765567136
Semantic Scholar
0728437b5553338297a6cc1bad9f15766a1e1c43
Cited by
1,338
References
53
Citations by year
2026: 110 · 2025: 129 · 2024: 143 · 2023: 139 · 2022: 166 · 2021: 181 · 2020: 190 · 2019: 170 · 2018: 100 · 2017: 6 · 2014: 1 · 2012: 1
FWCI
52.01
Type
journal article
Language
en
Open access
Licence
Locations
  • Cell (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • HAL (submittedVersion, cc-by)
Funders
  • European Research Council (232947)
  • Israeli Science Foundation
  • CNRS
  • FP7 European Network of Excellence EpiGeneSys (676556)
  • Agence Nationale de la Recherche (ANR-15-CE12-006-01)
  • Fondation pour la Recherche Médicale (144186)
  • INSERM and ITMO Cancer (PLBIO16-22)
  • Laboratory of Excellence (ANR-10-LABX-12-01)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Chromatin Remodeling and Cancer
  • Life Sciences › Agricultural and Biological Sciences › Plant Science › Chromosomal and Genetic Variations
Keywords
3D genome organization, chromatin conformation, neuronal differentiation, topologically associating domains, long-range chromatin interactions, chromatin insulation, enhancer-promoter interactions
MeSH terms
CCCTC-Binding Factor, Animals, Chromatin, Enhancer Elements, Genetic, Exons, Promoter Regions, Genetic, Repressor Proteins, Transcription Factors, Gene Expression, Genome, Mice, Gene Regulatory Networks, Embryonic Stem Cells, Neurogenesis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review