Chromatin architecture has been implicated in cell-type-specific gene regulatory programs; yet, how chromatin remodels during development remains to be fully elucidated. Here, by interrogating chromatin reorganization during human pluripotent stem cell (PSC) differentiation, we discover a role for the primate-specific endogenous retrotransposon HERV-H in creating topologically associating domains (TAD) in human PSCs. Deleting these HERV-H elements eliminates their corresponding TAD boundaries and reduces transcription of upstream genes, while de novo insertion of HERV-Hs can introduce new TAD boundaries. HERV-H’s ability to create these TAD boundaries depends on high transcription, as transcriptional repression of HERV-H elements prevents formation of these boundaries. This ability is not limited to human PSCs, as these actively transcribed HERV-Hs and their corresponding TAD boundaries also appear in PSCs from other hominids but not in more distantly related species lacking HERV-Hs. Overall, our results provide direct evidence for retrotransposons in actively shaping cell-type- and species-specific chromatin architecture.

ARID1A, a subunit of the SWI/SNF chromatin remodeling complex, is frequently mutated in cancer. Deficiency in its homolog ARID1B is synthetically lethal with ARID1A mutation. However, the functional relationship between these homologs has not been explored. Here, we use ATAC-seq, genome-wide histone modification mapping, and expression analysis to examine colorectal cancer cells lacking one or both ARID proteins. We find that ARID1A has a dominant role in maintaining chromatin accessibility at enhancers, while the contribution of ARID1B is evident only in the context of ARID1A mutation. Changes in accessibility are predictive of changes in expression and correlate with loss of H3K4me and H3K27ac marks, nucleosome spacing, and transcription factor binding, particularly at growth pathway genes including MET. We find that ARID1B knockdown in ARID1A mutant ovarian cancer cells causes similar loss of enhancer architecture, suggesting that this is a conserved function underlying the synthetic lethality between ARID1A and ARID1B.

Loss of heterochromatin has been implicated as a cause of pre-mature aging and age-associated decline in organ functions in mammals; however, the specific cell types and gene loci affected by this type of epigenetic change have remained unclear. To address this knowledge gap, we probed chromatin accessibility at single-cell resolution in the brains, hearts, skeletal muscles, and bone marrows from young, middle-aged, and old mice, and assessed age-associated changes at 353,126 candidate cis-regulatory elements (cCREs) across 32 major cell types. Unexpectedly, we detected increased chromatin accessibility within specific heterochromatin domains in old mouse excitatory neurons. The gain of chromatin accessibility at these genomic loci was accompanied by the cell-type-specific loss of heterochromatin and activation of LINE1 elements. Immunostaining further confirmed the loss of the heterochromatin mark H3K9me3 in the excitatory neurons but not in inhibitory neurons or glial cells. Our results reveal the cell-type-specific changes in chromatin landscapes in old mice and shed light on the scope of heterochromatin loss in mammalian aging.

BackgroundMicroarray experiments comprise more than half of all series in the Gene Expression Omnibus (GEO). However, downloading and analyzing raw or semi-processed microarray data from GEO is not intuitive and requires manual error-prone analysis and a bioinformatics background. This is due to a lack of standardization in array platform fabrication as well as the lack of a simple interactive tool for clustering, plotting, differential expression testing, and testing for functional enrichment.ResultsWe introduce the Bioinformatics Array Research Tool (BART), an R Shiny web application that automates the microarray download and analysis process across diverse microarray platforms. It provides an intuitive interface, automatically downloads and parses data from GEO, suggests groupings of samples for differential expression testing, performs batch effect correction, outputs quality control plots, converts probe IDs, generates full lists of differentially expressed genes, and performs functional enrichment analysis. We show that BART enables a more comprehensive analysis of a wider range of microarray datasets on GEO by comparing it to four leading online microarray analysis tools.ConclusionsBART allows a scientist with no bioinformatics background to extract knowledge from their own microarray data or microarray experiments available from GEO. BART is functional on more microarray experiments and provides more comprehensive analyses than extant microarray analysis tools. BART is hosted on bart.salk.edu, includes a user tutorial, and is available for download from https://bitbucket.org/Luisa_amaral/bart.

No abstract