Biology subjects

Tsirigos, A.

Publications and source records attributed to Tsirigos, A..

At least 55 records · Page 3Linked to original sources

bioRxiv2020

A CRISPR Screen Identifies Myc-associated Zinc Finger Protein (MAZ) as an Insulator Functioning at CTCF boundaries in Hox Clusters

The essential CCCTC-binding factor (CTCF) is critical to three-dimensional (3D) genome organization. CTCF binding insulates active and repressed genes within the Hox clusters upon differentiation, but such binding does not participate in boundary formation in all cell types, such as embryonic stem cells. We conducted a genome-wide CRISPR knockout screen to identify genes required for CTCF boundary activity at the HoxA cluster, complemented by novel biochemical approaches. This screen identified Myc-associated zinc finger protein (MAZ) as a CTCF insulator co-factor, among other candidates listed herein. MAZ depletion or specific deletion of MAZ motifs within the Hox clusters led to de-repression of posterior Hox genes immediately after CTCF boundaries upon differentiation, which phenocopied deletion of the proximal CTCF motifs. Similar to CTCF, MAZ interacted with the cohesin subunit, RAD21. Upon loss of MAZ, local contacts within topologically associated domains (TADs) were disrupted, as evidenced by HiC analysis. Thus, MAZ is a novel factor sharing insulation properties with CTCF and contributing to the genomic architectural organization. One Sentence SummaryMAZ is identified as an insulator functioning at CTCF boundaries delimiting active and repressed genes at Hox clusters

biochemistry↗

bioRxiv2020

Mitotic retention of H3K27 acetylation promotes rapid topological and transcriptional resetting of stem cell-related genes and enhancers upon G1 entry

The identity of dividing cells is challenged during mitosis, as transcription is halted and chromatin architecture drastically altered. How cell type-specific gene expression and genomic organization are faithfully reset upon G1 entry in daughter cells remains elusive. To address this issue, we characterized at a genome-wide scale the dynamic transcriptional and architectural resetting of mouse pluripotent stem cells (PSCs) upon mitotic exit. This revealed distinct patterns of transcriptional reactivation with rapid induction of stem cell genes and their enhancers, a more gradual recovery of metabolic and cell cycle genes, and a weak and transient activation of lineage-specific genes only during G1. Topological reorganization also occurred in an asynchronous manner and associated with the levels and kinetics of transcriptional reactivation. Chromatin interactions around active promoters and enhancers, and particularly super enhancers, reformed at a faster rate than CTCF/Cohesin-bound structural loops. Interestingly, regions with mitotic retention of the active histone mark H3K27ac and/or specific DNA binding factors showed faster transcriptional and architectural resetting, and chemical inhibition of H3K27 acetylation specifically during mitosis abrogated rapid reactivation of H3K27ac-bookmarked genes. Finally, we observed a contact between the promoter of an endoderm master regulator, Gata6, and a novel enhancer which was preestablished in PSCs and preserved during mitosis. Our study provides an integrative map of the topological and transcriptional changes that lead to the resetting of pluripotent stem cell identity during mitotic exit, and reveals distinct patterns and features that balance the dual requirements for self-renewal and differentiation.

molecular biology↗

bioRxiv2020

A bipartite element with allele-specific functions safeguards DNA methylation imprints at the Dlk1-Dio3 locus

Dysregulation of imprinted gene loci also referred to as loss of imprinting (LOI) can result in severe developmental defects and other diseases, but the molecular mechanisms that ensure imprint stability remain incompletely understood. Here, we dissect the functional components of the imprinting control region of the essential Dlk1-Dio3 locus (called IG-DMR) and the mechanism by which they ensure imprinting maintenance. Using pluripotent stem cells carrying an allele-specific reporter system, we demonstrate that the IG-DMR consists of two antagonistic regulatory elements: a paternally methylated CpG-island that prevents the activity of Tet dioxygenases and a maternally unmethylated regulatory element, which serves as a non-canonical enhancer and maintains expression of the maternal Gtl2 lncRNA by precluding de novo DNA methyltransferase function. Targeted genetic or epigenetic editing of these elements leads to LOI with either bi-paternal or bi-maternal expression patterns and respective allelic changes in DNA methylation and 3D chromatin topology of the entire Dlk1-Dio3 locus. Although the targeted repression of either IG-DMR or Gtl2 promoter is sufficient to cause LOI, the stability of LOI phenotype depends on the IG-DMR status, suggesting a functional hierarchy. These findings establish the IG-DMR as a novel type of bipartite control element and provide mechanistic insights into the control of Dlk1-Dio3 imprinting by allele-specific restriction of the DNA (de)methylation machinery. HIGHLIGHTSO_LIThe IG-DMR is a bipartite element with distinct allele-specific functions C_LIO_LIA non-canonical enhancer within the IG-DMR prevents DNA methyltransferase activity C_LIO_LITargeted epigenome editing allows induction of specific imprinting phenotypes C_LIO_LICRISPRi reveals a functional hierarchy between DMRs that dictates imprint stability C_LI

developmental biology↗

bioRxiv2020

Graph Drawing-based Dimensionality Reduction to Identify Hidden Communities in Single-Cell Sequencing Spatial Representation

With the rapid growth of single cell sequencing technologies, finding cell communities with high accuracy has become crucial for large scale projects. Employing the current commonly used dimensionality reduction techniques such as tSNE and UMAP, it is often difficult to clearly distinguish cell communities in high dimensional space. Usually cell communities with similar origin and trajectories cluster so closely to each that their subtle but important differences do not become readily apparent. This creates a problem for clustering, as clustering is also performed on dimensionality reduction results. In order to identify such communities, scientists either perform broad clustering and then extract each cluster and perform re-clustering to identify sub-populations or they over-cluster the data and then merging the clusters with similar gene expressions. This is an incredibly cumbersome and time-consuming process. To solve this problem, we propose K-nearest-neighbor-based Network graph drawing Layout (KNetL, pronounced like nettle) for dimensionality reduction. In our method, we use force-directed graph drawing, whereby the attractive force (analogous to a spring force) and the repulsive force (analogous to an electrical force in atomic particles) between the cells are evaluated, and the cell communities are organized in a structural visualization. The coordinates of the force-compacted nodes are then extracted, and we employ dimensionality reduction methods, such as tSNE and UMAP to unpack the nodes. The final plot, a KNetL map, shows a visually-appealing and distinctive separation between cell communities. Our results show that KNetL maps bring significant resolution to visualizing and identifying otherwise hidden cell communities. All the algorithms are implemented in the iCellR package and available through the CRAN repository. Single (i) Cell R package (iCellR) provides great flexibility at every step of the analysis pipeline, including normalization, clustering, dimensionality reduction, interactive 2D and 3D visualizations, batch alignment or data integration, imputation, and interactive cell gating tools, which allow users to manually gate around the cells.

bioinformatics↗

bioRxiv2020

iCellR: Combined Coverage Correction and Principal Component Alignment for Batch Alignment in Single-Cell Sequencing Analysis

Under-sampling RNA molecules and low-coverage sequencing in some single cell sequencing technologies introduce zero counts (also known as drop-outs) into the expression matrices. This issue may complicate the processes of dimensionality reduction and clustering, often forcing distinct cell types to falsely resemble one another, while eliminating subtle, but important differences. Considering the wide range in drop-out rates from different sequencing technologies, it can also affect the analysis at the time of batch/sample alignment and other downstream analyses. Therefore, generating an additional harmonized gene expression matrix is important. To address this, we introduce two separate batch alignment methods: Combined Coverage Correction Alignment (CCCA) and Combined Principal Component Alignment (CPCA). The first method uses a coverage correction approach (analogous to imputation) in a combined or joint fashion between multiple samples for batch alignment, while also correcting for drop-outs in a harmonious way. The second method (CPCA) skips the coverage correction step and uses k nearest neighbors (KNN) for aligning the PCs from the nearest neighboring cells in multiple samples. Our results of nine scRNA-seq PBMC samples from different batches and technologies shows the effectiveness of both these methods. All of our algorithms are implemented in R, deposited into CRAN, and available in the iCellR package.

bioinformatics↗

bioRxiv2019

Defining the relative and combined contribution of CTCF and CTCFL to genomic regulation

BackgroundUbiquitously expressed CTCF is involved in numerous cellular functions, such as organizing chromatin into TAD structures. In contrast, its paralog, CTCFL is normally only present in testis. However, it is also aberrantly expressed in many cancers. While it is known that shared and unique zinc finger sequences in CTCF and CTCFL enable CTCFL to bind competitively to a subset of CTCF binding sites as well as its own unique locations, the impact of CTCFL on chromosome organization and gene expression has not been comprehensively analyzed in the context of CTCF function. Using an inducible complementation system, we analyze the impact of expressing CTCFL and CTCF-CTCFL chimeric proteins in the presence or absence of endogenous CTCF to clarify the relative and combined contribution of CTCF and CTCFL to chromosome organization and transcription. ResultsWe demonstrate that the N terminus of CTCF interacts with cohesin which explains the requirement for convergent CTCF binding sites in loop formation. By analyzing CTCF and CTCFL binding in tandem we identify phenotypically distinct sites with respect to motifs, targeting to promoter/intronic intergenic regions and chromatin folding. Finally, we reveal that the N, C and zinc finger terminal domains play unique roles in targeting each paralog to distinct binding sites, to regulate transcription, chromatin looping and insulation. ConclusionThis study clarifies the unique and combined contribution of CTCF and CTCFL to chromosome organization and transcription, with direct implications for understanding how their co-expression deregulates transcription in cancer.

genomics↗

bioRxiv2019

Dynamic 3D chromosomal landscapes in acute leukemia

Three-dimensional (3D) chromatin architectural changes can alter the integrity of topologically associated domains (TADs) and rewire specific enhancer-promoter interactions impacting gene expression. Recently, such alterations have been implicated in human disease, highlighting the need for a deeper understanding of their role. Here, we investigate the reorganization of chromatin architecture in T cell acute lymphoblastic leukemia (T-ALL) using primary human leukemia specimens and its dynamic responses to pharmacological agents. Systematic integration of matched in situ Hi-C, RNA-Seq and CTCF ChIP-Seq datasets revealed widespread changes in intra-TAD chromatin interactions and TAD boundary insulation in T-ALL. Our studies identify and focus on a TAD \"fusion\" event being associated with loss of CTCF-mediated insulation, enabling direct interactions between the MYC promoter and a distal super-enhancer. Moreover, our data show that small molecule inhibitors targeting either oncogenic signal transduction or epigenetic regulation reduce specific 3D interactions associated with transformation. Overall, our study highlights the impact, complexity and dynamic nature of 3D chromatin architecture in human acute leukemia.\n\nOne Sentence Summary3D chromatin alterations in T cell leukemia are accompanied by changes in insulation and oncogene expression and can be partially restored by targeted drug treatments.

cancer biology↗

bioRxiv2019

Efficient pan-cancer whole-slide image classification and outlier detection using convolutional neural networks

Visual analysis of solid tissue mounted on glass slides is currently the primary method used by pathologists for determining the stage, type and subtypes of cancer. Although whole slide images are usually large (10s to 100s thousands pixels wide), an exhaustive though time-consuming assessment is necessary to reduce the risk of misdiagnosis. In an effort to address the many diagnostic challenges faced by trained experts, recent research has been focused on developing automatic prediction systems for this multi-class classification problem. Typically, complex convolutional neural network (CNN) architectures, such as Googles Inception, are used to tackle this problem. Here, we introduce a greatly simplified CNN architecture, PathCNN, which allows for more efficient use of computational resources and better classification performance. Using this improved architecture, we trained simultaneously on whole-slide images from multiple tumor sites and corresponding non-neoplastic tissue. Dimensionality reduction analysis of the weights of the last layer of the network capture groups of images that faithfully represent the different types of cancer, highlighting at the same time differences in staining and capturing outliers, artifacts and misclassification errors. Our code is available online at: https://github.com/sedab/PathCNN.

pathology↗

bioRxiv2019

Context-dependent requirement of H3K9 methyltransferase activity during cellular reprogramming to iPSCs

Methylation of histone 3 at lysine 9 (H3K9) is widely regarded as a major roadblock for cellular reprogramming and interference with associated methyltransferases such as EHMT1 and EHMT2 (also known as GLP and G9A, respectively) increases the efficiencies at which induced pluripotent stem cells (iPSCs) can be derived. Activation of histone and DNA demethylases by ascorbic acid (AA) has become a common approach to facilitate the extensive epigenetic remodeling required for iPSC formation, but possible functional interactions between the H3K9 methylation machinery and AA-stimulated enzymes remain insufficiently explored. Here we show that reduction of EHMT1/2 activity counteracts iPSC formation in an optimized reprogramming system in the presence of AA. Mechanistically, EHMT1/2 activity under these conditions is required for efficient downregulation of somatic genes and transition into an epithelial state. Of note, transient inhibition of EHMT1/2 during reprogramming yields iPSCs that fail to efficiently give rise to viable mice, suggesting persistent molecular defects in these cells. Genetic interference with the H3K9 demethylase KDM3B ameliorated the adverse effect of EHMT1/2 inhibition on iPSC formation. Together, our observations document novel functions of H3K9 methyltransferases during iPSC formation and suggest that the balancing of AA-stimulated enzymes by EHMT1/2 supports efficient and error-free iPSC reprogramming to pluripotency.

developmental biology↗

bioRxiv2019

A Deep Learning Approach for Rapid Mutational Screening in Melanoma

Image-based analysis as a rapid method for mutation detection can be advantageous in research or clinical settings when tumor tissue is limited or unavailable for direct testing. Here, we applied a deep convolutional neural network (CNN) to whole slide images of melanomas from 256 patients and developed a fully automated model that first selects for tumor-rich areas (Area Under the Curve AUC=0.96) then predicts for the presence of mutated BRAF in our test set (AUC=0.72) Model performance was cross-validated on melanoma images from The Cancer Genome Atlas (AUC=0.75). We confirm that the mutated BRAF genotype is linked to phenotypic alterations at the level of the nucleus through saliency mapping and pathomics analysis, which reveal that cells with mutated BRAF exhibit larger and rounder nuclei. Not only do these findings provide additional insights on how BRAF mutations affects tumor structural characteristics, deep learning-based analysis of histopathology images have the potential to be integrated into higher order models for understanding tumor biology, developing biomarkers, and predicting clinical outcomes.

pathology↗

bioRxiv2019

Gibbs Process Determines Survival and Reveals Contact-Inhibition Genes in Glioblastoma Multiforme

Tumor growth is a spatiotemporal birth-and-death process with loss of heterotypic contact-inhibition of locomotion (CIL) of tumor cells promoting invasion and metastasis. Therefore, representing tumor cells as two-dimensional points, we can expect the tumor tissues in histology slides to reflect realizations of spatial birth-and-death process which can be mathematically modeled to reveal molecular mechanisms of CIL, provided the mathematics models the inhibitory interactions. Gibbs process as an inhibitory point process is a natural choice since it is an equilibrium process of the spatial birth-and-death process. That is if the tumor cells maintain homotypic contact inhibition, the spatial distributions of tumor cells will result in Gibbs hard core process over long time scales. In order to verify if this is the case, we applied the Gibbs process to 411 TCGA Glioblastoma multiforme patient images. Our imaging dataset included all cases for which diagnostic slide images were available. The model revealed two clusters, one of which - the "Gibbs cluster," showed the convergence of the Gibbs process with significant survival difference. Further smoothing the discretized (and noisy) inhibition metric, for both increasing and randomized survival time, we found a significant association of the patients in the Gibbs cluster with increasing survival time. The mean inhibition metric also revealed the point at which the homotypic CIL establishes in tumor cells. Besides, RNAseq analysis between patients with loss of heterotypic CIL and intact homotypic CIL in the Gibbs cluster unveiled cell movement gene signatures and differences in Actin cytoskeleton and RhoA signaling pathways as key molecular alterations. These genes and pathways have established roles in CIL. Taken together, our integrated analysis of patient images and RNAseq data provides for the first time a mathematical basis for CIL in tumors, explains survival as well as uncovers the underlying molecular landscape for this key tumor invasion and metastatic phenomenon.

cancer biology↗

bioRxiv2019

NSD2 overexpression drives clustered chromatin and transcriptional changes in insulated domains

CTCF and cohesin play a key role in organizing chromatin into TAD structures. Disruption of a single CTCF binding site is sufficient to change chromosomal interactions leading to alterations in chromatin modifications and gene regulation. However, the extent to which alterations in chromatin modifications can disrupt 3D chromosome organization leading to transcriptional changes is unknown. In multiple myeloma a 4;14 translocation induces overexpression of the histone methyltransferase, NSD2 resulting in expansion of H3K36me2 and shrinkage of antagonistic H3K27me3 domains. Using isogenic cell lines producing high and low levels of NSD2, we find oncogene activation is linked to alterations in H3K27ac and CTCF within H3K36me2 enriched chromatin. A linear regression model reveals that changes in both CTCF and/or H3K27ac significantly increase the probability that a gene sharing the same insulated domain will be differentially expressed. These results identify a bidirectional relationship between 2D chromatin and 3D genome organization in gene regulation.

genomics↗

bioRxiv2018

The translatome of adult cortical axons is regulated by learning in vivo

Local translation can support memory consolidation by supplying new proteins to synapses undergoing plasticity. Translation in adult forebrain dendrites is an established mechanism of synaptic plasticity and is regulated by learning, yet there is no evidence for learning-regulated protein synthesis in adult forebrain axons, which have traditionally been believed to be incapable of translation. Here we show that axons in the adult rat amygdala contain translation machinery, and use translating ribosome affinity purification (TRAP) with RNASeq to identify mRNAs in cortical axons projecting to the amygdala, over 1200 of which were regulated during consolidation of associative memory. Mitochondrial and translation-related genes were upregulated, whereas synaptic, cytoskeletal, and myelin-related genes were downregulated; the opposite effects were observed in the cortex. Our results demonstrate that learning-regulated axonal translation occurs in the adult forebrain, and support the likelihood that local translation is more a rule than an exception in neuronal processes.

neuroscience↗

bioRxiv2018

KLF4 binding during reprogramming is involved in 3D architectural rewiring and transcriptional regulation of enhancer hubs

Cell fate transitions are accompanied by global transcriptional, epigenetic and topological changes driven by transcription factors (TFs), as is strikingly exemplified by reprogramming somatic cells to pluripotent stem cells (PSCs) via expression of OCT4, KLF4, SOX2 and cMYC. How TFs orchestrate the complex molecular changes around their target gene loci in a temporal manner remains incompletely understood. Here, using KLF4 as a paradigm, we provide the first TF-centric view of chromatin reorganization and its association to 3D enhancer rewiring and transcriptional changes of linked genes during reprogramming of mouse embryonic fibroblasts (MEFs) to PSCs. Inducible depletion of KLF factors in PSCs caused a genome-wide decrease in the connectivity of enhancers, while disruption of individual KLF4 binding sites from PSC-specific enhancers was sufficient to impair enhancer-promoter contacts and reduce expression of associated genes. Our study provides an integrative view of the complex activities of a lineage-specifying TF during a controlled cell fate transition and offers novel insights into the order and nature of molecular events that follow TF binding.

molecular biology↗

bioRxiv2018

ChIPSeqSpike: A R/Bioconductor package for ChIP-Seq data scaling according to spike-in control

MotivationChromatin Immuno-Precipitation followed by Sequencing (ChlP-Seq) is used to determine the binding sites of any protein of interest. ChIP-Seq data suffer from being more qualitative than quantitative. The recent use of Spike-in controls along with the standard protocol tackled this problem. However, no dedicated tool is available for a robust evaluation of this new ChIP-seq approach\n\nResultsWe developed ChIPSeqSpike, an R/Bioconductor package that enables ChIP-Seq spike-in normalization, assessment and analysis. Ready to use scaled bigwig files and scaling factors values are obtained as output. ChIPSeqSpike also provides tools for ChIP-Seq spike-in assessment and analysis through a versatile collection of graphical functions.\n\nAvailabilityThe package is implemented in R (as of version 3.4) and is available from Bioconductor at the URL: https://www.bioconductor.org/packages/3.7/bioc/html/ChIPSeqSpike.html, where installation and usage instructions can be found.\n\nContactnicolas.descostes@nyumc.org

bioinformatics↗

bioRxiv2017

Classification and Mutation Prediction from Non-Small Cell Lung Cancer Histopathology Images using Deep Learning

Visual analysis of histopathology slides of lung cell tissues is one of the main methods used by pathologists to assess the stage, types and sub-types of lung cancers. Adenocarcinoma and squamous cell carcinoma are two most prevalent sub-types of lung cancer, but their distinction can be challenging and time-consuming even for the expert eye. In this study, we trained a deep learning convolutional neural network (CNN) model (inception v3) on histopathology images obtained from The Cancer Genome Atlas (TCGA) to accurately classify whole-slide pathology images into adenocarcinoma, squamous cell carcinoma or normal lung tissue. Our method slightly outperforms a human pathologist, achieving better sensitivity and specificity, with [~]0.97 average Area Under the Curve (AUC) on a held-out population of whole-slide scans. Furthermore, we trained the neural network to predict the ten most commonly mutated genes in lung adenocarcinoma. We found that six of these genes - STK11, EGFR, FAT1, SETBP1, KRAS and TP53 - can be predicted from pathology images with an accuracy ranging from 0.733 to 0.856, as measured by the AUC on the held-out population. These findings suggest that deep learning models can offer both specialists and patients a fast, accurate and inexpensive detection of cancer types or gene mutations, and thus have a significant impact on cancer treatment.

cancer biology↗

bioRxiv2017

Robust Estimation Of Hi-C Contact Matrices Using Fused Lasso Reveals Preferential Insulation Of Super-Enhancers By Strong TAD Boundaries

The metazoan genome is compartmentalized in megabase-scale areas of highly interacting chromatin known as topologically associating domains (TADs), typically identified by computational analyses of Hi-C sequencing data. TADs are demarcated by boundaries that are largely conserved across cell types and even across species, although, increasing evidence suggests that the seemingly invariant TAD boundaries may exhibit plasticity and their insulating strength can vary. However, a genome-wide characterization of TAD boundary strength in mammals is still lacking. A systematic classification and characterization of TAD boundaries may generate new insights into their function. In this study, we first use fused two-dimensional lasso as a machine learning method to improve Hi-C contact matrix reproducibility, and, subsequently, we categorize TAD boundaries based on their insulation score. We demonstrate that higher TAD boundary insulation scores are associated with elevated CTCF levels and that they may differ across cell types. Intriguingly, we observe that super-enhancer elements are preferentially insulated by strong boundaries, i.e. boundaries of higher insulation score. Furthermore, we perform a pan-cancer analysis to demonstrate that strong TAD boundaries and super-enhancer elements are frequently co-duplicated in cancer patients. Taken together, our findings suggest that super-enhancers insulated by strong TAD boundaries may be exploited, as a functional unit, by cancer cells to promote oncogenesis.

bioinformatics↗

bioRxiv2016

lncRNA-screen: an interactive platform for computationally screening long non-coding RNAs in large genomics datasets

Long non-coding RNAs (lncRNAs) have emerged as a class of factors that are important for regulating development and cancer. Computational prediction of lncRNAs from ultra-deep RNA sequencing has been successful in identifying candidate lncRNAs. However, the complexity of handling and integrating different types of genomics data poses significant challenges to experimental laboratories that lack extensive genomics expertise. To address this issue, we have developed lncRNA-screen, a comprehensive pipeline for computationally screening putative lncRNA transcripts over large multimodal datasets. The main objective of this work is to facilitate the computational discovery of lncRNA candidates to be further examined by functional experiments. lncRNA-screen provides a fully automated easy-to-run pipeline which performs data download, RNA-seq alignment, assembly, quality assessment, transcript filtration, novel lncRNA identification, coding potential estimation, expression level quantification, histone mark enrichment profile integration, differential expression analysis, annotation with other type of segmented data (CNVs, SNPs, Hi-C, etc.) and visualization. Importantly, lncRNA-screen generates an interactive report summarizing all interesting lncRNA features including genome browser snapshots and lncRNA-mRNA interactions based on Hi-C data. In summary, our pipeline provides a comprehensive solution for lncRNA discovery and an intuitive interactive report for identifying promising lncRNA candidates. lncRNA-screen is available as free open-source software on GitHub.

bioinformatics↗