Biology subjects

Lee, A. J.

Publications and source records attributed to Lee, A. J..

At least 37 records · Page 2Linked to original sources

bioRxiv2024

Brain implantation of tissue-level-soft bioelectronics 1 via embryonic development

The design of bioelectronics capable of stably tracking brain-wide, single-cell, and millisecond-resolved neural activities in the developing brain is critical to the study of neuroscience and neurodevelopmental disorders. During development, the three-dimensional (3D) structure of the vertebrate brain arises from a 2D neural plate1,2. These large morphological changes previously posed a challenge for implantable bioelectronics to track neural activity throughout brain development3-9. Here, we present a tissue-level-soft, sub-micrometer-thick, stretchable mesh microelectrode array capable of integrating into the embryonic neural plate of vertebrates by leveraging the 2D-to-3D reconfiguration process of the tissue itself. Driven by the expansion and folding processes of organogenesis, the stretchable mesh electrode array deforms, stretches, and distributes throughout the entire brain, fully integrating into the 3D tissue structure. Immunostaining, gene expression analysis, and behavioral testing show no discernible impact on brain development or function. The embedded electrode array enables long-term, stable, brain-wide, single-unit-single-spike-resolved electrical mapping throughout brain development, illustrating how neural electrical activities and population dynamics emerge and evolve during brain development.

bioengineering↗

bioRxiv2024

Data-driven fine-grained region discovery in the mouse brain with transformers

Spatial transcriptomics offers unique opportunities to define the spatial organization of tissues and organs, such as the mouse brain. We address a key bottleneck in the analysis of organ-scale spatial transcriptomic data by establishing a workflow for self-supervised spatial domain detection that is scalable to multimillion cell datasets. This workflow uses a self-supervised framework for learning latent representations of tissue spatial domains or niches. We use a novel encoder-decoder architecture, which we named CellTransformer, to hierarchically learn higher-order tissue features from lower-level cellular and molecular statistical patterns. Coupling our representation learning workflow with minibatched GPU-accelerated clustering algorithms allows us to scale to multi-million cell MERFISH datasets where other methods cannot. CellTransformer is effective at integrating cells across tissue sections, identifying domains highly similar to ones in existing ontologies such as Allen Mouse Brain Common Coordinate Framework (CCF) while allowing discovery of hundreds of uncataloged areas with minimal loss of domain spatial coherence. CellTransformer domains recapitulate previous neuroanatomical studies of areas in the subiculum and superior colliculus, and characterize putatively uncataloged subregions in subcortical areas which currently lack subregion annotation. CellTransformer is also capable of domain discovery in whole-brain Slide-seqV2 datasets. Our workflows enable complex multi-animal analyses, achieving nearly perfect consistency of up to 100 spatial domains in a dataset of four individual mice with nine million cells across more than 200 tissue sections. CellTransformer advances the state of the art for spatial transcriptomics, by providing a performant solution for detection of fine-grained tissue domains from spatial transcriptomics data.

bioinformatics↗

bioRxiv2024

An image segmentation pipeline optimized for human microglia uncovers sources of morphological diversity in Alzheimer's disease

Alzheimers disease (AD) is a neurodegenerative disorder characterized by cognitive decline and the accumulation of beta-amyloid plaques and neurofibrillary tangles. Microglia, the resident immune cells in the brain, play a crucial role in AD pathology, particularly in relation to amyloid plaques. However, the exact role of plaque-associated microglia and their morphological changes in AD progression remains under debate. In this study, we aimed to establish an automated image segmentation and analysis pipeline optimized for the detection of human microglia, and we assessed its utility by systematically investigating topological relationships between microglia morphology and amyloid plaques. We accessed post-mortem brain tissue samples from persons with AD and utilized immunofluorescence staining to label microglia (IBA1) and amyloid plaques (A{beta}1-42). The acquired images were processed using CellProfiler, an open-source image analysis software, to automatically segment microglia and measure their morphological features. Specifically, since activated (stage III) microglia have a condensed morphology with retracted processes, we prioritized a morphological parameter called "compactness" in our analyses to capture shape changes in microglial found in proximity to amyloid plaques. Our results revealed that microglia unassociated with plaques (Mip-) were more abundant than microglia associated with plaques (Mip+) in the Dorsolateral Prefrontal Cortex (DLPFC) of both men and women with AD. Furthermore, we observed that Mip+ exhibited a significantly more ramified shape and had a higher expression of the IBA1 microglial marker gene compared to Mip. There were no significant differences in microglia morphology between men and women. Our study highlights the utility of automated image analysis in characterizing detailed microglia morphology at the single-cell level and its relationship with AD pathology.

pathology↗

bioRxiv2023

Atrophy-driven functional network collapse in neurodegenerative disease

Cognitive and behavioral deficits in Alzheimers disease (AD) and frontotemporal dementia (FTD) result from brain atrophy and altered functional connectivity. However, it is unclear how atrophy relates to functional connectivity disruptions across dementia subtypes and stages. We addressed this question using structural and functional MRI from 221 patients with AD (n=82), behavioral variant FTD (n=41), corticobasal syndrome (n=27), nonfluent (n=34) and semantic (n=37) variant primary progressive aphasia, and 100 cognitively normal individuals. Using partial least squares regression, we identified three principal structure-function components. The first component showed overall atrophy correlating with primary cortical hypo-connectivity and subcortical/association cortical hyper-connectivity. Components two and three linked focal syndrome-specific atrophy to peri-lesional hypo-connectivity and distal hyper-connectivity. Structural and functional component scores predicted global and domain-specific cognitive deficits. Anatomically, functional connectivity changes reflected alterations in specific brain activity gradients. Eigenmode analysis identified temporal phase and amplitude collapse as an explanation for atrophy-driven functional connectivity changes.

neuroscience↗

bioRxiv2023

Psychosocial experiences are associated with human brain mitochondrial biology

Psychosocial experiences affect brain health and aging trajectories, but the molecular pathways underlying these associations remain unclear. Normal brain function relies heavily on energy transformation by mitochondria oxidative phosphorylation (OxPhos), and two main lines of evidence bi-directionally link mitochondria as both targets and drivers of psychosocial experiences. On the one hand, chronic stress exposure and possibly mood states alter multiple aspects of mitochondrial biology; and on the other hand, functional variations in mitochondrial OxPhos capacity alter social behavior, stress reactivity, and mood. However, knowledge on whether positive or negative psychosocial exposures and experiences are linked to mitochondrial biology in the human brain is currently unknown. By combining longitudinal antemortem assessments of psychosocial factors with postmortem brain (dorsolateral prefrontal cortex) proteomics in older adults, we found that positive experiences (e.g. higher well-being) are linked to greater abundance of the mitochondrial OxPhos machinery, whereas negative experiences (e.g. higher negative mood) are linked to lower OxPhos protein content. Combined, psychosocial factors explained 18% of the variance in the abundance of OxPhos complex I, the primary biochemical entry point that energizes brain mitochondria. To increase the sensitivity of our approach, we next interrogated mitochondrial psychobiological associations in specific neuronal and non-neuronal brain cells with single-nucleus RNA sequencing. These results revealed strong cell type specific associations, particularly between positive psychosocial experiences and molecular mitochondrial phenotypes in glial cells, whereas neurons tended to show opposite associations. Accordingly, in bulk transcriptomic analyses where all cells are pooled, these RNA-based associations were masked. Thus, our results highlight the likely underestimation of effect sizes in bulk brain tissues, and document novel cell type specific mitochondrial psychobiological associations in the human brain. Cell type specific mitochondrial recalibrations represent a potential psychobiological pathway linking positive and negative psychosocial experiences to human brain biology. Significance statementPsychosocial experiences predict health trajectories, but the underlying mechanism remains unclear. We found that positive psychosocial experiences are linked to greater abundance of the mitochondrial energy transformation machinery, whereas negative experiences are linked to lower abundance. Overall, we found that psychosocial experiences explain 18% of the variance in abundance of complex I proteins, the main entry point of the mitochondrial oxidative phosphorylation (OxPhos) system. At single-cell resolution using single nucleus transcriptomics, positive psychosocial experiences were particularly related to glial cell mitochondrial phenotypes. Opposite associations between glial cells and neurons were naturally masked in bulk transcriptomic analyses. Our results suggest that mitochondrial recalibrations in specific brain cell types may represent a potential psychobiological pathway linking psychosocial experiences to human brain health.

cell biology↗

bioRxiv2023

The subacromial bursa is a key regulator of the rotator cuff and a new therapeutic target for improving repair

Rotator cuff injuries result in over 500,000 surgeries performed annually, an alarmingly high number of which fail. These procedures typically involve repair of the injured tendon and removal of the subacromial bursa. However, recent identification of a resident population of mesenchymal stem cells and inflammatory responsiveness of the bursa to tendinopathy indicate an unexplored biological role of the bursa in the context of rotator cuff disease. Therefore, we aimed to understand the clinical relevance of bursa-tendon crosstalk, characterize the biologic role of the bursa within the shoulder, and test the therapeutic potential for targeting the bursa. Proteomic profiling of patient bursa and tendon samples demonstrated that the bursa is activated by tendon injury. Using a rat to model rotator cuff injury and repair, tenotomy-activated bursa protected the intact tendon adjacent to the injured tendon and maintained the morphology of the underlying bone. The bursa also promoted an early inflammatory response in the injured tendon, initiating key players in wound healing. In vivo results were supported by targeted organ culture studies of the bursa. To examine the potential to therapeutically target the bursa, dexamethasone was delivered to the bursa, prompting a shift in cellular signaling towards resolution of inflammation in the healing tendon. In conclusion, contrary to current clinical practice, the bursa should be retained to the greatest extent possible and provides a new therapeutically target for improving tendon healing outcomes. One Sentence SummaryThe subacromial bursa is activated by rotator cuff injury and regulates the paracrine environment of the shoulder to maintain the properties of the underlying tendon and bone.

bioengineering↗

bioRxiv2023

ZCCHC17 modulates neuronal RNA splicing and supports cognitive resilience in Alzheimer's disease

ZCCHC17 is a putative master regulator of synaptic gene dysfunction in Alzheimers Disease (AD), and ZCCHC17 protein declines early in AD brain tissue, before significant gliosis or neuronal loss. Here, we investigate the function of ZCCHC17 and its role in AD pathogenesis. Co-immunoprecipitation of ZCCHC17 followed by mass spectrometry analysis in human iPSC-derived neurons reveals that ZCCHC17s binding partners are enriched for RNA splicing proteins. ZCCHC17 knockdown results in widespread RNA splicing changes that significantly overlap with splicing changes found in AD brain tissue, with synaptic genes commonly affected. ZCCHC17 expression correlates with cognitive resilience in AD patients, and we uncover an APOE4 dependent negative correlation of ZCCHC17 expression with tangle burden. Furthermore, a majority of ZCCHC17 interactors also co-IP with known tau interactors, and we find significant overlap between alternatively spliced genes in ZCCHC17 knockdown and tau overexpression neurons. These results demonstrate ZCCHC17s role in neuronal RNA processing and its interaction with pathology and cognitive resilience in AD, and suggest that maintenance of ZCCHC17 function may be a therapeutic strategy for preserving cognitive function in the setting of AD pathology. SignificanceAbnormal RNA processing is an important component of AD pathophysiology. We show here that ZCCHC17, a previously identified putative master regulator of synaptic dysfunction in AD, plays a role in neuronal RNA processing, and illustrate that ZCCHC17 dysfunction is sufficient to explain some of the splicing abnormalities seen in AD brain tissue, including synaptic gene splicing abnormalities. Using data from human patients, we demonstrate that ZCCHC17 mRNA levels correlate with cognitive resilience in the setting of AD pathology. These results suggest that maintenance of ZCCHC17 function may be a therapeutic strategy for supporting cognitive function in AD patients, and motivate future work examining a possible role of abnormal RNA processing in AD-related cognitive decline.

neuroscience↗

bioRxiv2023

Multi-region brain transcriptomes uncover two subtypes of aging individuals with differences in Alzheimer risk and the impact of APOEe4

The heterogeneity of the older population suggests the existence of subsets of individuals which share certain brain molecular features and respond differently to risk factors for Alzheimers disease, but this population structure remains poorly defined. Here, we performed an unsupervised clustering of individuals with multi-region brain transcriptomes to assess whether a broader approach, simultaneously considering data from multiple regions involved in cognition would uncover such subsets. We implemented a canonical correlation-based analysis in a Discovery cohort of 459 participants from two longitudinal studies of cognitive aging that have RNA sequence profiles in three brain regions. 690 additional participants that have data in only one or two of these regions were used in the Replication effort. These clustering analyses identified two meta-clusters, MC-1 and MC-2. The two sets of participants differ primarily in their trajectories of cognitive decline, with MC-2 having a delay of 3 years to the median age of incident dementia. This is due, in part, to a greater impact of tau pathology on neuronal chromatin architecture and to broader brain changes including greater loss of white matter integrity in MC-1. Further evidence of biological differences includes a significantly larger impact of APOE{varepsilon}4 risk on cognitive decline in MC-1. These findings suggest that our proposed population structure captures an aspect of the more distributed molecular state of the aging brain that either enhances the effect of risk factors in MC-1 or of protective effects in MC-2. These observations may inform the design of therapeutic development efforts and of trials as both become increasingly more targeted molecularly. One Sentence Summary: There are two types of aging brains, with one being more vulnerable to APOE{varepsilon}4 and subsequent neuronal dysfunction and cognitive loss.

neuroscience↗

bioRxiv2023

Volumetric refractive index measurement and quantitative density analysis of mouse brain tissue with sub-micrometer spatial resolution

High-resolution structural imaging of brain tissue is important for neuroscience research. However, conventional approaches have several limitations, such as the need for exogenous staining, limited accessibility to volumetric information, and qualitative analysis. Herein, we present high-resolution label-free volumetric imaging and analysis of mouse brain tissue using three-dimensional quantitative phase imaging. Measurement of the refractive index distribution of tissue enables direct imaging of the cellular and subcellular structures. Quantification of subcellular organelles is performed in the anatomical regions of the somatosensory cortex, corpus callosum, caudoputamen, and thalamus regions.

biophysics↗

bioRxiv2022

A cross-disease human microglial framework identifies disease-enriched subsets and tool compounds for microglial polarization.

Human microglia play a pivotal role in neurological diseases, but few targeted therapies that directly modulate microglial state or function exist due to an incomplete understanding of microglial heterogeneity. We use single-cell RNA sequencing to profile live human microglia from autopsies or surgical resections across diverse neurological diseases and central nervous system regions. We observe a central divide between oxidative and heterocyclic metabolism and identify subsets associated with antigen presentation, motility, and proliferation. Specific subsets are enriched in susceptibility genes for neurodegenerative diseases or the disease-associated microglial signature. We validate subtypes in situ with an RNAscope-immunofluorescence pipeline and leverage our dataset as a classification resource, finding that iPSC model systems recapitulate substantial in vivo heterogeneity. Finally, we identify and validate candidates for chemically inducing subtype-specific states in vitro, showing that Camptothecin downregulates the transcriptional signature of disease-enriched subsets and upregulates a signature previously shown to be depleted in Alzheimers.

neuroscience↗

bioRxiv2022

Compendium-wide analysis of P. aeruginosa core and accessory genes reveal more nuanced transcriptional patterns

Strains of Pseudomonas aeruginosa, an opportunistic pathogen that causes difficult to treat infections, have significant genomic heterogeneity including the presence of diverse accessory genes that are only present in some strains or clades. Both core genes, which are conserved across strains, and accessory genes have been associated with traits such as biofilm formation and virulence. Much of what we know about core and accessory gene content comes from genome analyses. Here, we use a newly assembled transcriptome compendium to analyze the transcriptional patterns of core and accessory gene expression in PAO1 and PA14 strains across thousands of samples from hundreds of distinct experiments. We found that a subset of core genes were stable, having consistent correlated expression patterns across samples regardless of strain background, with a focus on strains PAO1 and PA14. These stable core genes had fewer co-expressed neighbors that were accessory genes. SignificancePseudomonas aeruginosa is a ubiquitous pathogen. There is a lot of diversity amongst P. aeruginosa strains, some which are clinically relevant. Understanding how these different strain-level traits manifest is important for identifying targets that regulate different traits of interest. With the availability of a PAO1-mapped and PA14-mapped RNA-seq compendium, which contain hundreds of strains, it is now possible to examine the effect of different strains on expression, which can mediate different traits. In this study we developed an approach to compare expression profiles across different P. aeruginosa gene groups - core and accessory genes. This approach revealed a subset of core genes with different transcriptional patterns across strains, which could contribute to trait differences.

bioinformatics↗

bioRxiv2022

CHA1: A New Combinatorial Therapy That Reciprocally Regulates Wnt and JAK/STAT/Interferon Signaling to Re-program Breast Tumors and the Tumor-Resident Landscape.

Triple negative breast cancers (TNBC) pose exceptional challenges with fatal brain metastases as a clear and unmet need. Immune checkpoint inhibitors (ICIs) are promising therapeutic strategies, but most TNBC are resistant, or "cold" tumors, due to lack of tumor-resident immune engagement. No FDA-approved therapies exist which promote a "cold-to-hot" transition or induce the important biomarker PD-L1, often used for ICI clinical decision-making. Maximal ICI susceptibility, or a full "cold-to-hot" transition, requires reciprocal Wnt signaling inhibition and Jak/STAT/interferon signaling activation. We report a new compound combination (CHA1) that fits the above criteria. CHA1 is comprised of EGCG (epigallocatechin-3-gallate; green-tea compound) and decitabine (DNA-methyltransferase (DNMT1) inhibitor; 5deaza-cytidine; FDA-approved for hematologic malignancies). We used immune-compromised and syngeneic TNBC pre-clinical models to investigate tumor-intrinsic and tumor-resident T-cell effects, respectively. All results required CHA1 (but not EGCG or decitabine alone) and utilized attainable human dose equivalences with manageable safety profiles. CHA1 triggered efficient Wnt signaling inhibition by elevating Wnt pathway inhibitors (HBP1 and SFRP1) and traversed the blood-brain barrier to reduce both tumor and brain metastatic growth. Transcriptomic and expression analyses revealed that CHA1 treatment effectuated a robust tumor-intrinsic JAK/STAT/IFN response 1) to induce PDL1 and 2) to induce antigen presentation and processing genes, including MHC-1, MHC-2 and numerous genes attributed to professional antigen-presenting cells; 3) to induce CD8+-T-cell infiltration and activation. Additionally, CHA1 pre-treatment improved anti-PDL1 efficacy in a syngeneic setting. Lastly, we derived a composite gene signature emblematic of CHA1 treatment and of a favorable clinical prognosis in-silico. Together, our work supports a model in which CHA1 influences epigenetics, Wnt and Jak/STAT/IFN signaling mechanisms--all to reprogram an epithelial-mesenchymal TNBC tumor to express antigen-presenting properties and to recruit and activate tumor-resident CD8+-T-cells. We discuss our findings in the context of cancer biology and immunity with implications for improving ICI susceptibility for TNBC.

cancer biology↗

bioRxiv2022

CF-Seq, An Accessible Web Application for Rapid Re-Analysis of Cystic Fibrosis Pathogen RNA Sequencing Studies

Researchers studying cystic fibrosis (CF) pathogens have produced numerous RNA-seq datasets which are available in the gene expression omnibus (GEO). Although these studies are publicly available, substantial computational expertise and manual effort are required to compare similar studies, visualize gene expression patterns within studies, and use published data to generate new experimental hypotheses. Furthermore, it is difficult to filter available studies by domain-relevant attributes such as strain, treatment, or media, or for a researcher to assess how a specific gene responds to various experimental conditions across studies. To reduce these barriers to data re-analysis, we have developed an R Shiny application called CF-Seq, which works with a compendium of 147 studies and 1,446 individual samples from 13 clinically relevant CF pathogens. The application allows users to filter studies by experimental factors and to view complex differential gene expression analyses at the click of a button. Here we present a series of use cases that demonstrate the application is a useful and efficient tool for new hypothesis generation. (CFSeq: http://scangeo.dartmouth.edu/CFSeq/)

bioinformatics↗

bioRxiv2022

Computationally efficient assembly of a Pseudomonas aeruginosa gene expression compendium

Over the past two decades, thousands of RNA sequencing (RNA-seq) gene expression profiles of Pseudomonas aeruginosa have been made publicly available via the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA). In the work we present here, we draw on over 2,300 P. aeruginosa transcriptomes from hundreds of studies performed by over seventy-five different research groups. We first developed a pipeline, using the Salmon pseudo-aligner and two different P. aeruginosa reference genomes (strains PAO1 and PA14), that transformed raw sequence data into a uniformly processed data in the form of sample-wise normalized counts. In this workflow, P. aeruginosa RNA-seq data are filtered using technically and biologically driven criteria with characteristics tailored to bacterial gene expression and that account for the effects of alignment to different reference genomes. The filtered data are then normalized to enable cross experiment comparisons. Finally, annotations are programmatically collected for those samples with sufficient meta-data and expression-based metrics are used to further enhance strain assignment for each sample. Our processing and quality control methods provide a scalable framework for taking full advantage of the troves of biological information hibernating in the depths of microbial gene expression data. The re-analysis of these data in aggregate is a powerful approach for hypothesis generation and testing, and this approach can be applied to transcriptome datasets in other species. SignificancePseudomonas aeruginosa causes a wide range of infections including chronic infections associated with cystic fibrosis. P. aeruginosa infections are difficult to treat and people with CF-associated P. aeruginosa infections often have poor clinical outcomes. To aid the study of this important pathogen, we developed a methodology that facilitates analyses across experiments, strains, and conditions. We aligned, filtered for quality and normalized thousands of P. aeruginosa RNA-seq gene expression profiles that were publicly available via the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA). The workflow that we present can be efficiently scaled to incorporate new data and applied to the analysis of other species.

microbiology↗

bioRxiv2021

Label-free monitoring of 3D cortical neuronal growth in vitro using optical diffraction tomography

The highly complex central nervous systems of mammals are often studied using three-dimensional (3D) in vitro primary neuronal cultures. A coupled confocal microscopy and immunofluorescence labeling are widely utilized for visualizing the 3D structures of neurons. However, this requires fixation of the neurons and is not suitable for monitoring an identical sample at multiple time points. Thus, we propose a label-free monitoring method for 3D neuronal growth based on refractive index tomograms obtained by optical diffraction tomography. The 3D morphology of the neurons was clearly visualized, and the developmental processes of neurite outgrowth in 3D spaces were analyzed for individual neurons. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=101 SRC="FIGDIR/small/454602v1_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@b858baorg.highwire.dtl.DTLVardef@c946caorg.highwire.dtl.DTLVardef@468168org.highwire.dtl.DTLVardef@da022c_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2021

Exocyst Inactivation in Urothelial Cells Disrupts Autophagy and Activates non-canonical NF-κB

Ureter obstruction is a highly prevalent event during embryonic development and is a major cause of pediatric kidney disease. We have reported that ureteric bud-specific ablation of the exocyst Exoc5 subunit in late-murine gestation results in failure of urothelial stratification, cell death, and complete ureter obstruction. However, the mechanistic connection between disrupted exocyst activity, urothelial cell death, and subsequent ureter obstruction was unclear. Here, we report that inhibited urothelial stratification does not drive cell death during ureter development. Instead, we demonstrate that the exocyst plays a critical role in autophagy in urothelial cells, and that disruption of autophagy activates a urothelial NF-{kappa}B stress response. Impaired autophagy first provokes canonical NF-{kappa}B activity which is progressively followed by increasing non-canonical NF-{kappa}B activity and cell death if the stress remains unresolved. Furthermore, we demonstrate that ureter obstructions can be completely rescued in Exoc5 conditional knockout mice by administering a single dose of pan-caspase inhibitor z-VAD-FMK at E16.5 prior to urothelial cell death. Taken together, ablation of Exoc5 disrupts autophagic stress response and activates progressive NF-{kappa}B signaling which promotes obstructive uropathy.

developmental biology↗

bioRxiv2021

CHD4 conceals aberrant CTCF-binding sites at TAD interiors by regulating chromatin accessibility in mESCs

CTCF plays a critical role in the 3D chromatin organization by determining the TAD borders. Although CTCF primarily binds at the TAD borders, there also exist putative CTCF-binding sites within TADs, which are spread throughout the genome by retrotransposition. However, the detailed mechanism responsible for masking these putative CTCF-binding sites remains elusive. Here, we show that the ATP-dependent chromatin remodeler, CHD4, regulates chromatin accessibility to conceal aberrant CTCF-binding sites embedded in H3K9me3-enriched heterochromatic B2 SINEs in mouse embryonic stem cells (mESCs). Upon CHD4 depletion, these aberrant CTCF-binding sites become accessible, and aberrant CTCF recruitment occurs at the TAD interiors, resulting in disorganization of the local TADs. Furthermore, RNA-binding intrinsically disordered domains of CHD4 is required to prevent the aberrant CTCF bindings. Lastly, CHD4 is required for the repression of B2 SINE transcripts. These results highlight the CHD4-mediated mechanism that safeguards the appropriate CTCF bindings and associated TAD organizations in mESCs.

molecular biology↗

bioRxiv2021

Generative neural networks separate common and specific transcriptional responses

Genome-wide transcriptome profiling identifies genes that are prone to differential expression across contexts ("common DEGs"), as well as genes with changes specific to the experimental manipulation. Distinguishing common DEGs from those that are specifically changed in a context of interest allows more efficient prediction of which genes are specific to a given biological process under scrutiny. Currently, commonly differentially expressed genes or pathways can only be identified through the laborious manual curation of highly controlled experiments, an inordinately time-consuming and impractical endeavor. Here we pioneer an approach for identifying common patterns using generative neural networks. This approach produces a background set of transcriptomic experiments from which a null distribution of gene and pathway changes can be generated. By comparing the set of differentially expressed genes found in a target experiment against the generated background set, common results can be easily separated from specific ones. This "Specific cOntext Pattern Highlighting In Expression data" (SOPHIE) approach is broadly applicable to new platforms or any species with a large collection of gene expression data. We apply SOPHIE to diverse datasets including those from human, human cancer, and the bacterial pathogen Pseudomonas aeruginosa. SOPHIE identifies common DEGs in concordance with previously described, manually and systematically determined common DEGs. Further, molecular validation indicates that SOPHIE detects highly specific, but low magnitude, biologically relevant, transcriptional changes. SOPHIEs measure of specificity can complement log fold change values generated from traditional differential expression analyses. For example, by filtering the set of differentially expressed genes, one can identify those genes that are specifically relevant to the experimental condition of interest. Consequently, these results can inform future research directions.

bioinformatics↗