An intricate molecular machinery is at the core of regulation of gene expression within every cell. During the initial stages of organismal development, the coordinated activation of diverse transcriptional programs is crucial and must be carefully executed to shape every organ and tissue. Bivalent promoters and poised enhancers are regulatory regions decorated with histone marks associated with both positive and negative transcriptional outcomes. These apparently contradictory signals are important for setting bivalent genes in a poised state, which is subsequently resolved during differentiation into either active or repressive conditions. Here, we discuss the origins of bivalent promoters and the mechanisms implicated in their acquisition and maintenance. We further review how the presence of bivalent marks influences genome architecture.Finally, we highlight the potential link between bivalency and cancer which could drive biomedical research in disease etiology and treatment.

The ChIP-seq signal of histone modifications at promoters is a good predictor of gene expression in different cellular contexts, but whether this is also true at enhancers is not clear. To address this issue, we develop quantitative models to characterize the relationship of gene expression with histone modifications at enhancers or promoters. We use embryonic stem cells (ESCs), which contain a full spectrum of active and repressed (poised) enhancers, to train predictive models. As many poised enhancers in ESCs switch towards an active state during differentiation, predictive models can also be trained on poised enhancers throughout differentiation and in development. Remarkably, we determine that histone modifications at enhancers, as well as promoters, are predictive of gene expression in ESCs and throughout differentiation and development. Importantly, we demonstrate that their contribution to the predictive models varies depending on their location in enhancers or promoters. Moreover, we use a local regression (LOESS) to normalize sequencing data from different sources, which allows us to apply predictive models trained in a specific cellular context to a different one. We conclude that the relationship between gene expression and histone modifications at enhancers is universal and different from promoters. Our study provides new insight into how histone modifications relate to gene expression based on their location in enhancers or promoters.

Context & Objective Chronic glucocorticoid (GC) overexposure, resulting from endogenous Cushing’s syndrome (CS) or exogenous GC therapy, causes several adverse outcomes, including persistent central fat accumulation associated with a low-grade inflammation. However, no previous multi-omics studies in visceral adipose tissue (VAT) from patients exposed to high levels of unsuppressed GC during active CS or after remission are available yet. Methods We employed a translational approach combining high-throughput data on endogenous CS patients and a reversible CS mouse model. We performed RNA-seq and ChIP-seq on histone modifications (H3K4me3, H3K27ac and H3K27me3) to identify persistent transcriptional and epigenetic signatures in VAT produced during active CS and maintained after remission. Results VAT dysfunction was associated with low-grade pro-inflammatory status, macrophage infiltration and extracellular matrix remodeling. Most notably, chronic hypercortisolism caused a persistent circadian rhythm disruption in VAT through core clock genes modulation. Importantly, changes in the levels of two histone modifications associated to gene transcriptional activation (H3K4me3 and H3K27ac) correlated with the observed differences in gene expression during active CS and after CS remission. Conclusion We identified for the first time, the persistent transcriptional and epigenetic signatures induced by hypercortisolism in VAT, providing a novel integrated view of molecular components driving the long-term VAT impairment associated with CS.

Large-scale sequencing techniques to chart genomes are entirely consolidated. Stable computational methods to perform primary tasks such as quality control, read mapping, peak calling, and counting are likewise available. However, there is a lack of uniform standards for graphical data mining, which is also of central importance. To fill this gap, we developed SeqCode, an open suite of applications that analyzes sequencing data in an elegant but efficient manner. Our software is a portable resource written in ANSI C that can be expected to work for almost all genomes in any computational configuration. Furthermore, we offer a user-friendly front-end web server that integrates SeqCode functions with other graphical analysis tools. Our analysis and visualization toolkit represents a significant improvement in terms of performance and usability as compare to other existing programs. Thus, SeqCode has the potential to become a key multipurpose instrument for high-throughput professional analysis; further, it provides an extremely useful open educational platform for the world-wide scientific community. SeqCode website is hosted at http://ldicrocelab.crg.eu, and the source code is freely distributed at https://github.com/eblancoga/seqcode.

No abstract