Background: Chromosomes are subdivided spatially to delimit long-range interactions into topologically associating domains (TADs). TADs are often flanked by chromatin insulators and transcription units that may participate in such demarcation. Remarkably, single-cell Drosophila TAD units correspond to dynamic heterochromatin nanocompartments that can self-assemble. The influence of insulators on such dynamic compartmentalization remains unclear. Moreover, to what extent heterochromatin domains are fully compartmentalized away from active genes remains unclear from Drosophila to human. Results: Here, we identify H3K27me3 micro-domains genome-wide in Drosophila, which are attributed to the three-dimensional spreading of heterochromatin marks into euchromatin. Whereas depletion of insulator proteins increases H3K27me3 spreading locally, across heterochromatin borders, it concomitantly decreases H3K27me3 levels at distant micro-domains discrete sites. Quantifying long-range interactions suggests that random interactions between heterochromatin TADs and neighbor euchromatin cannot predict the presence of micro-domains, arguing against the hypothesis that they reflect defects in self-folding or in insulating repressive TADs. Rather, microdomains are predicted by specific long-range interactions with the TAD borders bound by insulator proteins and co-factors required for looping. Accordingly, H3K27me3 spreading to distant sites is impaired by insulator mutants that compromise recruitment of looping co-factors. Both depletions and insulator mutants significantly reduce H3K27me3 micro-domains, deregulating the flanking genes. Conclusions: Our data highlight a new regulatory mode of H3K27me3 by insulatorbased long-range interactions controlling distant euchromatic genes.

BackgroundDNA inside eukaryotic cells wraps around histones to form the 11nm chromatin fiber that can further fold into higher-order DNA loops, which may depend on the binding of architectural factors. Predicting how the DNA will fold given a distribution of bound factors, here viewed as a type of sequence, is currently an unsolved problem and several heterogeneous polymer models have shown that many features of the measured structure can be reproduced from simulations. However a model that determines the optimal connection between sequence and structure and that can rapidly assess the effects of varying either one is still lacking.ResultsHere we train a dense neural network to solve for the local folding of chromatin, connecting structure, represented as a contact map, to a sequence of bound chromatin factors. The network includes a convolutional filter that compresses the large number of bound chromatin factors into a single 1D sequence representation that is optimized for predicting structure. We also train a network to solve the inverse problem, namely given only structural information in the form of a contact map, predict the likely sequence of chromatin states that generated it.ConclusionsBy carrying out sensitivity analysis on both networks, we are able to highlight the importance of chromatin contexts and neighborhoods for regulating long-range contacts, along with critical alterations that affect contact formation. Our analysis shows that the networks have learned physical insights that are informative and intuitive about this complex polymer problem.Electronic supplementary materialThe online version of this article (10.1186/s12859-018-2286-z) contains supplementary material, which is available to authorized users.

The histone variant macroH2A1.1 plays a role in cancer development and metastasis. To determine the underlying molecular mechanisms, we mapped the genome-wide localization of endogenous macroH2A1.1 in the human breast cancer cell line MDA-MB 231. We demonstrate that macroH2A1.1 specifically binds to active promoters and enhancers in addition to facultative heterochromatin. Selective knock-down of macroH2A1.1 deregulates expression of hundreds of highly active genes. Depending on the chromatin landscape, macroH2A1.1 acts through two distinct molecular mechanisms. The first mitigates excessive transcription by binding over domains including the promoter and the gene body. The second stimulates expression of RNA Polymerase II (Pol II) paused genes, including genes regulating mammary tumor cell migration. In contrast to the first one, macroH2A1.1 specifically associates with the TSS of Pol II paused genes. These processes occur in a predefined local 3D genome landscape but do not require rewiring of enhancer-promoter contacts. We thus propose that macroH2A1.1 serves as a transcriptional modulator with a potential role in assisting the conversion of promoter-locked into a productive and elongating Pol II.