[This corrects the article DOI: 10.1371/journal.pgen.1004604.].

With their ability to integrate their genetic material into the target cell genome, retroviral vectors (RV) of both the gamma-retroviral (γ-RV) and lentiviral vector (LV) classes currently remain the most efficient and thus the system of choice for achieving transgene retention and therefore potentially long-term expression and therapeutic benefit. However, γ-RV and LV integration comes at a cost in that transcription units will be present within a native chromatin environment and thus be subject to epigenetic effects (DNA methylation, histone modifications) that can negatively impact on their function. Indeed, highly variable expression and silencing of γ-RV and LV transgenes especially resulting from promoter DNA methylation is well documented and was the cause of the failure of gene therapy in a clinical trial for X-linked chronic granulomatous disease. This review will critically explore the use of different classes of genetic control elements that can in principle reduce vector insertion site position effects and epigenetic-mediated silencing. These transcriptional regulatory elements broadly divide themselves into either those with a chromatin boundary or border function (scaffold/matrix attachment regions, insulators) or those with a dominant chromatin remodeling and transcriptional activating capability (locus control regions,, ubiquitous chromatin opening elements). All these types of elements have their strengths and weaknesses within the constraints of a γ-RV and LV backbone, showing varying degrees of efficacy in improving reproducibility and stability of transgene function. Combinations of boundary and chromatin remodeling; transcriptional activating elements, which do not impede vector production; transduction efficiency; and stability are most likely to meet the requirements within a gene therapy context especially when targeting a stem cell population.

While sequencing-based proximity assays have established many megabase-scale features of chromosome organization, experimental noise and biases have left finer-scale organization poorly understood. We developed HiFive, an empirically-driven probabilistic modeling approach for normalizing and analyzing HiC and 5C data. HiFive allows reconstruction of chromatin structural features at higher resolution than previously described and meets or outperforms other approaches. With novel feature detection, denoising, and efficient algorithms, HiFive enables utilization of these data with a fraction of the computational power and time without sacrificing resolution. Using HiFive, we demonstrate transcriptionally-associated differential spatial organization of the mouse genome.

While a rich set of putative cis-regulatory sequences involved in mouse fetal development has been annotated recently based on chromatin accessibility and histone modification patterns, delineating their role in developmentally regulated gene expression continues to be challenging. To fill this gap, we mapped chromatin contacts between gene promoters and distal sequences genome-wide in seven mouse fetal tissues, and for one of them, across six developmental stages. We identified 248,620 long-range chromatin interactions centered at 14,138 protein-coding genes and characterized their tissue-to-tissue variations as well as developmental dynamics. Integrative analysis of the interactome with previous epigenome and transcriptome datasets from the same tissues revealed a strong correlation between the chromatin contacts and chromatin state at distal enhancers, as well as gene expression patterns at predicted target genes. We predicted target genes of 15,098 candidate enhancers, and used them to annotate target genes of homologous candidate enhancers in the human genome that harbor risk variants of human diseases. We present evidence that schizophrenia and other adult disease risk variants are frequently found in fetal enhancers, providing support for the hypothesis of fetal origins of adult diseases.

We review recent developments in mapping chromosomal contacts and compare emerging insights on chromosomal contact domain organization in Drosophila and mammalian cells. Potential scenarios leading to the observation of Hi-C domains and their association with the epigenomic context of the chromosomal elements involved are discussed. We argue that even though the mechanisms and precise physical structure underlying chromosomal domain demarcation are yet to be fully resolved, the implications to genome regulation and genome evolution are profound. Specifically, we hypothesize that domains are facilitating genomic compartmentalization that support the implementation of complex, modular, and tissue specific transcriptional program in metazoa.

We present a fast and simple algorithm to detect nascent RNA transcription in global nuclear run-on sequencing (GRO-seq). GRO-seq is a relatively new protocol that captures nascent transcripts from actively engaged polymerase, providing a direct read-out on bona fide transcription. Most traditional assays, such as RNA-seq, measure steady state RNA levels which are affected by transcription, post-transcriptional processing, and RNA stability. GRO-seq data, however, presents unique analysis challenges that are only beginning to be addressed. Here, we describe a new algorithm, Fast Read Stitcher (FStitch), that takes advantage of two popular machine-learning techniques, hidden Markov models and logistic regression, to classify which regions of the genome are transcribed. Given a small user-defined training set, our algorithm is accurate, robust to varying read depth, annotation agnostic, and fast. Analysis of GRO-seq data without a priori need for annotation uncovers surprising new insights into several aspects of the transcription process.

We investigate spatiotemporal dynamics of human interphase chromosomes by employing a heteropolymer model that incorporates the information of human chromosomes inferred from Hi-C data. Despite considerable heterogeneities in the chromosome structures generated from our model, chromatins are organized into crumpled globules with space-filling (SF) statistics characterized by a single universal scaling exponent (ν = 1/3), and this exponent alone can offer a quantitative account of experimentally observed, many different features of chromosome dynamics. The local chromosome structures, whose scale corresponds to that of topologically associated domains (∼ 0.1 − 1 Mb), display dynamics with a fast relaxation time (≲ 1 − 10 sec); in contrast, the long-range spatial reorganization of the entire chromatin (≳O(102) Mb) occurs on a much slower time scale (≳ hour), providing the dynamic basis of cell-to-cell variability and glass-like behavior of chromosomes. Biological activities, modeled using stronger isotropic white noises added to active loci, accelerate the relaxation dynamics of chromatin domains associated with the low frequency modes and induce phase segregation between the active and inactive loci. Surprisingly, however, they do not significantly change the dynamics at local scales from those obtained under passive conditions. Our study underscores the role of chain organization of chromosome in determining the spatiotemporal dynamics of chromatin loci.

We have developed a machine learning approach to predict stimulation-dependent enhancer-promoter interactions using evidence from changes in genomic protein occupancy over time. The occupancy of estrogen receptor alpha (ERα), RNA polymerase (Pol II) and histone marks H2AZ and H3K4me3 were measured over time using ChIP-Seq experiments in MCF7 cells stimulated with estrogen. A Bayesian classifier was developed which uses the correlation of temporal binding patterns at enhancers and promoters and genomic proximity as features to predict interactions. This method was trained using experimentally determined interactions from the same system and was shown to achieve much higher precision than predictions based on the genomic proximity of nearest ERα binding. We use the method to identify a genome-wide confident set of ERα target genes and their regulatory enhancers genome-wide. Validation with publicly available GRO-Seq data demonstrates that our predicted targets are much more likely to show early nascent transcription than predictions based on genomic ERα binding proximity alone.

We discuss here a series of testable hypotheses concerning the role of chromosome folding into topologically associating domains (TADs). Several lines of evidence suggest that segmental packaging of chromosomal neighborhoods may underlie features of chromatin that span large domains, such as heterochromatin blocks, association with the nuclear lamina and replication timing. By defining which DNA elements preferentially contact each other, the segmentation of chromosomes into TADs may also underlie many properties of long‐range transcriptional regulation. Several observations suggest that TADs can indeed provide a structural basis to regulatory landscapes, by controlling enhancer sharing and allocation. We also discuss how TADs may shape the evolution of chromosomes, by causing maintenance of synteny over large chromosomal segments. Finally we suggest a series of experiments to challenge these ideas and provide concrete examples illustrating how they could be practically applied.

We developed a targeted chromosome conformation capture (4C) approach that uses unique molecular identifiers (UMIs) to derive high-complexity quantitative chromosome contact profiles with controlled signal-to-noise ratios. UMI-4C detects chromosomal interactions with improved sensitivity and specificity, and it can easily be multiplexed to allow robust comparison of contact distributions between loci and conditions. This approach may open the way to the incorporation of contact distributions into quantitative models of gene regulation.

We describe an algorithm that can mine differential gene expression data to identify candidate cell type-specific DNA regulatory sequences. Differential expression is usually quantified as a continuous score—fold-change, test-statistic, p-value—comparing biological classes. Unlike existing approaches, our de novo strategy, termed SArKS, applies nonparametric kernel smoothing to uncover promoter motifs that correlate with elevated differential expression scores. SArKS detects motifs by smoothing sequence scores over sequence similarity. A second round of smoothing over spatial proximity reveals multimotif domains (MMDs). Discovered motifs can then be merged or extended based on adjacency within MMDs. False positive rates are estimated and controlled by permutation testing. We apply SArKS to a published gene expression data set representing distinct neocortical neuron classes in M. musculus. When benchmarked against several existing algorithms for correlative motif discovery using a cross-validation procedure, SArKS identified larger motif sets that formed the basis for regression models with higher correlative power. Finally, we examined specific motifs derived by SArKS from putative promoters of genes enriched in parvalbumin-expressing GABAergic interneurons, identifying several known biologically meaningful sequence elements.

We describe a Hi-C-based method, Micro-C, in which micrococcal nuclease is used instead of restriction enzymes to fragment chromatin, enabling nucleosome resolution chromosome folding maps. Analysis of Micro-C maps for budding yeast reveals abundant self-associating domains similar to those reported in other species, but not previously observed in yeast. These structures, far shorter than topologically associating domains in mammals, typically encompass one to five genes in yeast. Strong boundaries between self-associating domains occur at promoters of highly transcribed genes and regions of rapid histone turnover that are typically bound by the RSC chromatin-remodeling complex. Investigation of chromosome folding in mutants confirms roles for RSC, "gene looping" factor Ssu72, Mediator, H3K56 acetyltransferase Rtt109, and the N-terminal tail of H4 in folding of the yeast genome. This approach provides detailed structural maps of a eukaryotic genome, and our findings provide insights into the machinery underlying chromosome compaction.

We carried out an integrative analysis of enhancer landscape and gene expression dynamics during hematopoietic differentiation using DNase-seq, histone mark ChIP-seq and RNA sequencing to model how the early establishment of enhancers and regulatory locus complexity govern gene expression changes at cell state transitions. We found that high-complexity genes—those with a large total number of DNase-mapped enhancers across the lineage—differ architecturally and functionally from low-complexity genes, achieve larger expression changes and are enriched for both cell type–specific and transition enhancers, which are established in hematopoietic stem and progenitor cells and maintained in one differentiated cell fate but lost in others. We then developed a quantitative model to accurately predict gene expression changes from the DNA sequence content and lineage history of active enhancers. Our method suggests a new mechanistic role for PU.1 at transition peaks during B cell specification and can be used to correct assignments of enhancers to genes.

Understanding the genetic basis of human traits/diseases and the underlying mechanisms of how these traits/diseases are affected by genetic variations is critical for public health. Current genome-wide functional genomics data uncovered a large number of functional elements in the noncoding regions of human genome, providing new opportunities to study regulatory variants (RVs). RVs play important roles in transcription factor bindings, chromatin states and epigenetic modifications. Here, we systematically review an array of methods currently used to map RVs as well as the computational approaches in annotating and interpreting their regulatory effects, with emphasis on regulatory single-nucleotide polymorphism. We also briefly introduce experimental methods to validate these functional RVs.

Understanding the complexity of transcriptional regulation is a major goal of computational biology. Because experimental linkage of regulatory sites to genes is challenging, computational methods considering epigenomics data have been proposed to create tissue-specific regulatory maps. However, we showed that these approaches are not well suited to account for the variations of the regulatory landscape between cell-types. To overcome these drawbacks, we developed a new method called STITCHIT, that identifies and links putative regulatory sites to genes. Within STITCHIT, we consider the chromatin accessibility signal of all samples jointly to identify regions exhibiting a signal variation related to the expression of a distinct gene. STITCHIT outperforms previous approaches in various validation experiments and was used with a genome-wide CRISPR-Cas9 screen to prioritize novel doxorubicin-resistance genes and their associated non-coding regulatory regions. We believe that our work paves the way for a more refined understanding of transcriptional regulation at the gene-level.

Type 1 diabetes (T1D) is a chronic metabolic disorder characterized by the autoimmune destruction of insulin-producing pancreatic islet beta cells in genetically predisposed individuals. Genome-wide association studies (GWAS) have identified over 60 risk regions across the human genome, marked by single nucleotide polymorphisms (SNPs), which confer genetic predisposition to T1D. There is increasing evidence that disease-associated SNPs can alter gene expression through spatial interactions that involve distal loci, in a tissue- and development-specific manner. Here, we used three-dimensional (3D) genome organization data to identify genes that physically co-localized with DNA regions that contained T1D-associated SNPs in the nucleus. Analysis of these SNP-gene pairs using the Genotype-Tissue Expression database identified a subset of SNPs that significantly affected gene expression. We identified 246 spatially regulated genes including HLA-DRB1, LAT, MICA, BTN3A2, CTLA4, CD226, NOTCH1, TRIM26, PTEN, TYK2, CTSH, and FLRT3, which exhibit tissue-specific effects in multiple tissues. We observed that the T1D-associated variants interconnect through networks that form part of the immune regulatory pathways, including immune-cell activation, cytokine signaling, and programmed cell death protein-1 (PD-1). Our results implicate T1D-associated variants in tissue and cell-type specific regulatory networks that contribute to pancreatic beta cell inflammation and destruction, adaptive immune signaling, and immune-cell proliferation and activation. A number of other regulatory changes we identified are not typically considered to be central to the pathology of T1D. Collectively, our data represent a novel resource for the hypothesis-driven development of diagnostic, prognostic, and therapeutic interventions in T1D.

Transcriptional regulatory changes have been shown to contribute to phenotypic differences between species, but many questions remain about how gene expression evolves. Here we report the first comparative study of nascent transcription in primates. We used PRO-seq to map actively transcribing RNA polymerases in resting and activated CD4+ T-cells in multiple human, chimpanzee, and rhesus macaque individuals, with rodents as outgroups. This approach allowed us to measure transcription separately from post-transcriptional processes. We observed general conservation in coding and non-coding transcription, punctuated by numerous differences between species, particularly at distal enhancers and non-coding RNAs. We found evidence that transcription factor binding sites are a primary determinant of transcriptional differences between species, that stabilizing selection maintains gene expression levels despite frequent changes at distal enhancers, and that adaptive substitutions have driven lineage-specific transcription. Finally, we found strong correlations between evolutionary rates and long-range chromatin interactions. These observations clarify the role of primary transcription in regulatory evolution.

Transcriptional regulation is tightly coupled with chromosomal positioning and three-dimensional chromatin architecture. However, it is unclear what proportion of transcriptional activity is reflecting such organisation, how much can be informed by RNA expression alone, and how this impacts disease. Here, we develop a transcriptional decomposition approach separating the proportion of expression associated with genome organisation from independent effects not directly related to genomic positioning. We show that positionally attributable expression accounts for a considerable proportion of total levels and is highly informative of topological associating domain activities and organisation, revealing boundaries and chromatin compartments. Furthermore, expression data alone accurately predicts individual enhancer-promoter interactions, drawing features from expression strength, stabilities, insulation and distance. We further characterise commonalities and differences across predictions in 76 human cell types, observing extensive sharing of domains, yet highly cell-type specific enhancer-promoter interactions and strong enrichments in relevant trait-associated variants. Our work demonstrates a close relationship between transcription and chromatin architecture, presenting a novel strategy and an unprecedented resource for investigating regulatory organisations and interpretations of disease associated genetic variants across cell types.

Transcriptional enhancers regulate spatio-temporal gene expression. While genomic assays can identify putative enhancers en masse, assigning target genes is a complex challenge. We devised a machine learning approach, McEnhancer, which links target genes to putative enhancers via a semi-supervised learning algorithm that predicts gene expression patterns based on enriched sequence features. Predicted expression patterns were 73–98% accurate, predicted assignments showed strong Hi-C interaction enrichment, enhancer-associated histone modifications were evident, and known functional motifs were recovered. Our model provides a general framework to link globally identified enhancers to targets and contributes to deciphering the regulatory genome.

Tracking mitotic chromosome formation How cells pack DNA into fully compact, rod-shaped chromosomes during mitosis has fascinated cell biologists for more than a century. Gibcus et al. delineated the conformational transition trajectory from interphase chromatin to mitotic chromosomes minute by minute during the cell cycle. The mitotic chromosome is organized in a spiral staircase architecture in which chromatin loops emanate radially from a centrally located helical scaffold. The molecular machines condensin I and II play distinct roles in these processes: Condensin II is essential for helical winding, whereas condensin I modulates the organization within each helical turn. Science, this issue p. eaao6135 Mitotic chromosome folding involves formation of increasingly compacted helically arranged nested loop arrays. INTRODUCTION During mitosis, cells compact their chromosomes into dense rod-shaped structures to ensure their reliable transmission to daughter cells. Our work explores how cells achieve this compaction. We integrate genetic, genomic, and computational approaches to characterize the key steps in mitotic chromosome formation from the G2 nucleus to metaphase, and we identify roles of specific molecular machines, condensin I and II, in these major conformational transitions. RATIONALE We used chicken DT-40 cells expressing an analog-sensitive CDK1 to produce cell cultures that synchronously enter mitosis. We collected cells at key time points during mitotic entry; analyzed chromosome organization by microscopy, chromosome conformation capture, and polymer simulations; and delineated a pathway of mitotic chromosome formation. We used engineered cell lines to study the function of condensin complexes, which are critical for mitotic chromosome formation. We fused condensin I and II subunits to plant auxin-inducible degron domains, thus enabling their rapid depletion in late G2 just before mitotic entry. These cell lines allowed us to determine the roles of condensin I and II in specific steps of the mitotic chromosome morphogenesis pathway. RESULTS Our analysis of G2 chromosomes reveals hallmarks of interphase chromosomes, including topologically associating domains and compartments. Upon entry into prophase, this organization is lost within minutes, and by late prophase, chromosomes are folded as arrays of consecutive loops condensed around a central axis. These loops project with random but mutually correlated angles from the axis. During prometaphase, the loop array undergoes two major reorganizations. First, it acquires a helical arrangement of loops. Polymer simulations of Hi-C data show that the centrally located axis acquires a helical twist so that consecutive loops emanate as the steps of a spiral staircase. Second, the chromatin loops become nested with ~400-kb outer loops split up by ~80-kb inner loops. As prometaphase proceeds, chromosomes shorten through progressive helical winding, with the numbers of loops per turn increasing. As a result, the size of a helical turn grows from ~3 Mb (~40 loops) to ~12 Mb (~150 loops). Depletion of condensin I or II before mitotic entry revealed their differing roles in mitotic chromosome formation. Either condensin can mediate loop array formation. However, condensin II is required for the helical twisting of the scaffold from which loops emanate, whereas condensin I modulates the size and arrangement of nested inner loops. CONCLUSION We describe a pathway of mitotic chromosome folding that unifies many previous observations. In prophase, condensins mediate the loss of interphase organization and the formation of arrays of consecutive loops. In prometaphase, chromosomes adopt a spiral staircase–like structure with a helically arranged axial scaffold of condensin II at the bases of chromatin loops. The condensin II loops are further compacted by condensin I into clusters of smaller nested loops that are additionally collapsed by chromatin-to-chromatin attractions. The combination of nested loops distributed around a helically twisted axis plus dense chromatin packing achieves the 10,000-fold compaction of chromatin into linearly organized chromosomes that is required for accurate chromosome segregation when cells divide. A pathway for mitotic chromosome formation. In prophase, condensins mediate the loss of interphase chromosome conformation, and loop arrays are formed. In prometaphase, the combined action of condensin I (blue spheres in the bottom diagram) and II (red spheres) results in helically arranged nested loop arrays. Mitotic chromosomes fold as compact arrays of chromatin loops. To identify the pathway of mitotic chromosome formation, we combined imaging and Hi-C analysis of synchronous DT40 cell cultures with polymer simulations. Here we show that in prophase, the interphase organization is rapidly lost in a condensin-dependent manner, and arrays of consecutive 60-kilobase (kb) loops are formed. During prometaphase, ~80-kb inner loops are nested within ~400-kb outer loops. The loop array acquires a helical arrangement with consecutive loops emanating from a central “spiral staircase” condensin scaffold. The size of helical turns progressively increases to ~12 megabases during prometaphase. Acute depletion of condensin I or II shows that nested loops form by differential action of the two condensins, whereas condensin II is required for helical winding.