Biology subjects

Baek, S.

Publications and source records attributed to Baek, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Coordinated Temporal Dynamics of Glucocorticoid Receptor Binding and Chromatin Landscape Drive Transcriptional Regulation

Glucocorticoid receptor (GR) signaling elicits diverse transcriptional responses through dynamic and context-dependent interactions with chromatin. Here, we define a temporally resolved and mechanistically integrated framework for GR-mediated gene regulation. Time-resolved analyses identify three conserved classes of GR chromatin binding (sustained, transient, and late), distinguished by differences in motif strength, chromatin accessibility, and cofactors engagement. Early GR binding preferentially occurs at high-affinity glucocorticoid response elements (GREs) within pre-accessible regulatory regions, whereas late binding is associated with weaker motifs and requires chromatin remodeling activity. Enhancer activation, marked by H3K27ac deposition, closely tracks GR occupancy, supporting a model in which GR recruits acetyltransferase activity to drive coordinated enhancer activation. Concurrently, GR-centered interaction networks are dynamically reconfigured, and motif enrichment analyses identify distinct transcription factor signatures across binding classes, including AP-1/JUNB at transient sites and CEBP family members at late-binding regions. Integration of chromatin binding, chromatin interaction, and transcriptomic datasets reveals that temporal and combinatorial GR occupancy is functionally linked to gene expression programs. Distinct GR binding clusters are nonrandomly associated with specific transcriptional trajectories, including sustained, transient, and late gene induction. Moreover, combinatorial occupancy across multiple regulatory elements correlates quantitatively with transcriptional output, indicating that GR functions not as a simple binary regulator, but as an integrator of multilayered regulatory inputs. These findings support a unified model in which temporal binding dynamics, chromatin state, and combinatorial enhancer activity collectively encode transcriptional specificity, providing a general framework for stimulus-responsive nuclear receptor signaling.

molecular biology↗

bioRxiv2025

A Microstructurally-Motivated Framework to Study Autoregulation in the Coronary Circulation

Coronary autoregulation maintains relatively constant myocardial flow over a wide range of perfusion pressures through myogenic, shear-dependent, and metabolic control mechanisms. Understanding this phenomenon is challenging due to the coupled nature of these mechanisms and their heterogeneous effects throughout the coronary tree. In this study we developed a novel microstructurally-motivated model of coronary autoregulation based on constrained mixture theory, with anatomical and structural parameters calibrated through a homeostatic optimization framework. Autoregulation was simulated at three myocardial depths (subepicardium, midwall, and subendocardium), with the calibrated model accurately reproducing baseline hemodynamics and autoregulatory responses. For changes in epicardial pressure, our model reproduced experimentally measured subendocardium-to-subepicardium flow ratios (ENDO/EPI) and changes in vessel diameter, demonstrating its predictive capability. Furthermore, we extended Womersleys theory to simulate phasic coronary hemodynamics with a time-varying intramyocardial pressure. This microstructurally-motivated framework provides a mechanistic foundation for investigating coronary autoregulation and long-term vascular growth and remodeling in pathphysiological conditions. SummaryO_LICoronary autoregulation is defined as the capability of the coronary circulation to maintain the blood supply to the heart over a range of perfusion pressures. This phenomenon is facilitated through intrinsic mechanisms that control the vascular resistance by regulating the mechanical function of smooth muscle cells. Understanding the mechanisms involved in coronary autoregulation is one of the most fundamental questions in coronary physiology. C_LIO_LIThis paper presents a structurally-motivated coronary autoregulation model that uses a nonlinear continuum mechanics approach to account for the morphometry and vessel wall composition in two coronary trees in the subepicardial and subendocardial layers. C_LIO_LIThe model is calibrated against diverse experimental data from literature and is used to study heterogeneous autoregulatory response in the coronary trees. This model drastically differs from previous models, which relied on lumped parameter model formulations, and is suited to the study of long-term pathophysiological growth and remodeling phenomena in coronary vessels. C_LI

physiology↗

bioRxiv2025

Robust spatial hearing beyond primary interaural cues in humans over deep neural networks

Spatial hearing allows humans to localize sound sources in the azimuth plane using interaural time (ITD) and level (ILD) differences, but the contribution of additional auditory features remains unclear. To investigate this, we measured human localization performance with natural and artificial stimuli that selectively included or excluded ITD and ILD as primary interaural cues. As expected, human listeners relied synergistically on ITD and ILD for accurate azimuth localization. Moreover, even when both primary cues were absent, localization performance remained above chance level. We compared human performance with state-of-the-art deep neural networks (DNN) optimized for sound localization to investigate possible computational mechanisms underlying this robust performance. In contrast to humans, DNNs demonstrated high accuracy only for stimuli that resembled their training regime but failed when primary interaural cues were absent. This human-DNN misalignment highlights a fundamental distinction in sensory processing strategies, potentially arising from the simplicity bias inherent in DNN training, with human reliance on a wider range of auditory features likely reflecting evolutionary pressures favoring adaptability across diverse acoustic environments. Together, our results demonstrate the robustness of human spatial hearing beyond primary interaural cues and point to promising directions for advancing artificial systems and informing clinical applications, such as cochlear implants and auditory prosthetics.

animal behavior and cognition↗

bioRxiv2025

Linker Histone H1.5 Contributes to Centromere Integrity in Human Cells

Mammalian H1 linker histones comprise a group of 11 non-allelic variants which have key roles in modulating chromatin. H1 variant specific genomic distribution contributes to fine tuning regulation of gene expression and chromatin architecture. Contradictory reports on the presence and role of H1 histones at centromeres led us to further investigate whether H1s impact centromeric chromatin. In this study, we focused on H1.5 and by in vitro assays we showed that H1.5 directly interacts with centromeric-protein A (CENP-A) mononucleosomes. Notably, our in vitro findings revealed that H1 variants H1.0 and H1.2 can also bind CENP-A nucleosomes, although with differing affinities and signatures, asserting centromeric localization may not be unique to H1.5. In human cells, H1.5 localized to the centromere and chromatin immuno-precipitation revealed an interaction between H1.5 with CENP-A nucleosomes. Knocking down of H1.5 resulted in the loss of centromeric -satellite transcription, reduction in loading of new CENP-A, and the accumulation of mitotic defects. These data point to an unreported role for histone H1 in the regulation of mitotic integrity in human cells. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=120 SRC="FIGDIR/small/657682v1_ufig1.gif" ALT="Figure 1"> View larger version (29K): org.highwire.dtl.DTLVardef@26a35borg.highwire.dtl.DTLVardef@aecc44org.highwire.dtl.DTLVardef@1e49939org.highwire.dtl.DTLVardef@107a485_HPS_FORMAT_FIGEXP M_FIG Graphical AbstractProposed model of H1.5 function at the centromere. When H1.5 is bound to the CENP-A nucleosomes at the centromere, the centromeric chromatin allows RNAPII based transcription to generate -satellite transcripts, which in turn allows for de novo CENP-A loading. The absence of H1.5 blocks this transcription, leading to an accumulation of mitotic defects. C_FIG

biochemistry↗

bioRxiv2025

Oncogenic lncRNA transgene transcription modulates epigenetic memory at a naive chromosomal locus.

Maintaining genome integrity is crucial for the proper functioning and development of organisms. One intriguing aspect of genome integrity is the formation and function of neocentromeres at non-centromeric sites. CENP-A, a centromere-specific protein, is essential for centromere identification and function. However, in many cancers, CENP-A is often found to be ectopically misplaced when overexpressed. Moreover, CENP-A deposition at the centromere depends on the transcription of centromeric non-coding RNAs. Consequently, ectopic CENP-A is found at transcriptionally active and frequent breakpoint regions. To further explore ectopic CENP-A localization, we previously engineered a stable ectopic CENP-A site on a naive chromosome by overexpressing a non-centromeric oncogenic lncRNA, PCAT2, which was capable of recruiting CENP-A to its transcription site. In this work, we tracked cells carrying this stable transgene to understand the longevity of the induced ectopic CENP-A site at the chromosome that harbors it. Our findings revealed that the induced epigenetic memory was eventually lost due to the suppression of the transgene through competing epigenetic silencing mechanisms. This epigenetic restoration naturally reversed the ectopic CENP-A level to its previous levels at the engineered site. These data suggest that cells may have evolved failsafe mechanisms to prevent neocentromere formation at ectopic sites by suppressing transcription, unless otherwise favored by selection involving multiple components.

genetics↗

bioRxiv2025

Baseline State for Pulmonary Vasculature with Pulmonary Arterial Hypertension: Effect of Geometric Remodeling and Metabolic Shift

Pulmonary arterial hypertension (PAH) is a complex disease characterized by chronically elevated pulmonary arterial pressure, with early onset and progression linked to structural, metabolic and morphological changes in the pulmonary vasculature. Understanding the interplay between hemodynamics and arterial wall mechanics is essential to capture the pathology of the distal vasculature in PAH. This study aims to develop a data-driven framework that establishes a baseline state of PAH vasculature, incorporating key features of arterial wall constituents, geometry, and their interaction with PAH-specific hemodynamics. The model also explores changes in the metabolic energy costs of the arterial vasculature and hypothesize the most plausible metabolic costs based on PAH arterial wall energy consumption. Illustrative examples of symmetrically bifurcating arterial trees are used to establish baseline characteristics of PAH-affected pulmonary arteries. We establish the baseline state for PAH vasculature and compared it with healthy homeostatic vasculature in terms of arterial mechanics, morphometry, and pulsatile hemodynamics. This framework provides a representative computational model for advanced studies on PAH treatments and lays the groundwork for future pathophysiological modeling of the disease.

systems biology↗

bioRxiv2024

Augmenting the human interactome for disease prediction through gene networks inferred from human cell atlas

Gene co-expression network inference from bulk tissue samples often misses cell-type-specific interactions, which can be detected through single-cell gene expression data. However, the noise and sparsity of single-cell data challenge the inference of these networks. We developed scNET, a framework for integrative cell-type-specific co-expression network inference from single-cell transcriptome data, demonstrating its utility in augmenting the human interactome for more accurate disease gene prediction. We address the limitations in de novo network inference from single-cell expression data through dropout imputation, metacells formation, and data transformation. Employing this data preprocessing pipeline, we inferred cell-type-specific co-expression links from single-cell atlas data, covering various cell types and tissues, and integrated over 850K of these inferred links into a preexisting human interactome, HumanNet, resulting in HumanNet-plus. This integration notably enhanced accuracy of network-based disease gene prediction. These findings suggest that with proper data preprocessing, network inference from single-cell gene expression data can be highly effective, potentially enriching the human interactome and advancing the field of network medicine.

bioinformatics↗

bioRxiv2024

Hard-wired visual filters for environment-agnostic object recognition

Conventional deep neural networks (DNNs) are highly susceptible to variations in input domains, unlike biological brains which effectively adapt to environmental changes. Here, we demonstrate that hard-wired Gabor filters, replicating the structure of receptive fields in the brains early visual pathway, facilitate environment-agnostic object recognition without overfitting. Our approach involved fixing the pre-designed Gabor filters in the early layers of DNNs, preventing any alterations during training. Despite the restricted learning flexibility of this model, our networks maintained robust performance even under significant domain shifts, in contrast to conventional DNNs that typically fail in similar conditions. We found that our model effectively clustered identical "classes" across diverse domains, while conventional DNNs tend to cluster images by "domain" in the latent space. We observed that the fixed Gabor filters enabled networks to encode global shape information rather than local texture features, thereby mitigating the risk of overfitting. One sentence summaryHard-wired Gabor filters enable environment-agnostic object recognition without overfitting. Research HighlightsO_LIConventional deep neural networks (DNNs) are vulnerable to input domain variations C_LIO_LIHard-wired Gabor filters facilitate environment-agnostic object recognition C_LIO_LIFixed Gabor filters prevent overfitting and facilitate shape-based classifications C_LIO_LIOur model cluster identical "classes" while conventional DNNs cluster by "domain" C_LI

neuroscience↗

bioRxiv2024

Tumor microenvironment distinctions between esophageal cancer subtypes explain varied immunotherapy responses

Esophageal cancer comprises two main subtypes: esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC). Previous studies have revealed distinct genomic characteristics between these subtypes, with ESCC sharing similarities with head and neck squamous cell carcinoma (HNSCC), and EAC aligning with gastric adenocarcinoma (GAC). Additionally, recent immunotherapy clinical trials have shown a higher response rate in ESCC and HNSCC compared to EAC and GAC. However, how the tumor microenvironment contributes to these varied immunotherapy responses remains unclear. In this study, through a comparative analysis of single-cell tumor transcriptomes from 35 patients with ESCC, EAC, HNSCC, or GAC, we identified two groups with tumor microenvironment distinctions: ESCC and HNSCC versus EAC and GAC, consistent with their genomic classifications. Malignant epithelial cells displayed distinct separations based on histological origin. In the tumor immune microenvironment, we observed an enrichment of CXCL13+CD8+ T cells and CXCL9+CXCL10+ tumor-associated macrophages (TAMs) in ESCC and HNSCC, which activate cellular immunity through interferon-{gamma}. In contrast, EAC and GAC exhibited a high presence of heat-shock protein-expressing CD8+ T cells and MARCO+ TAMs. These immune signatures help explain the varied immunotherapy responses among these cancer subtypes and successfully predict immunotherapy outcomes across diverse cancer types, underscoring their clinical significance.

cancer biology↗

bioRxiv2024

Single-cell multi-omics reveals tumor microenvironment factors underlying poor immunotherapy responses in ALK-positive lung cancer

Anaplastic lymphoma kinase (ALK) rearrangement is a major oncogenic driver in non-small cell lung cancer (NSCLC). While ALK tyrosine kinase inhibitors have shown promising therapeutic effects, overcoming resistance with immunotherapy becomes necessary when resistance develops. However, various clinical trials have revealed that their efficacies remain limited. To investigate the tumor microenvironment (TME) factors contributing to poor immune checkpoint blockade responses in ALK-positive patients, we performed single-cell RNA and ATAC sequencing on lung adenocarcinoma (LUAD) tumors with and without ALK rearrangements. Integrative analysis with additional public LUAD cohorts revealed distinct immune landscapes in ALK-positive tumors, marked by enriched innate immunity and depleted adaptive immunity. ALK-positive malignant cells exhibit higher stemness and aggressive phenotype. Tumor-associated macrophages (TAMs) in these tumors predominantly maintain pro-tumoral M2-like states, reinforcing immune suppression. B cells show reduced immune reactivity and impaired tertiary lymphoid structure formation, while CD8+ T cells display bystander-like signatures and reduced tumor reactivity. Single-cell chromatin accessibility profiles combined with regulatory network analysis suggest that differences in transcription factor activities, rather than chromatin accessibility, may underlie T cell dysfunction. These findings provide insights into the immunosuppressive TME of ALK-positive LUAD, potentially explaining the failure of recent immunotherapy trials and highlighting targets for improving efficacy.

cancer biology↗

bioRxiv2024

Predictive Prioritization of Enhancers Associated with Pancreas Disease Risk

Genetic and epigenetic variations in regulatory enhancer elements increase susceptibility to a range of pathologies. Despite recent advances, linking enhancer elements to target genes and predicting transcriptional outcomes of enhancer dysfunction remain significant challenges. Using 3D chromatin conformation assays, we generated an extensive enhancer interaction dataset for the human pancreas, encompassing more than 20 donors and five major cell types, including both exocrine and endocrine compartments. We employed a network approach to parse chromatin interactions into enhancer-promoter tree models, facilitating a quantitative, genome-wide analysis of enhancer connectivity. With these tree models, we developed a machine learning algorithm to estimate the impact of enhancer perturbations on cell type-specific gene expression in the human pancreas. Orthogonal to our computational approach, we perturbed enhancer function in primary human pancreas cells using CRISPR interference and quantified the effects at the single-cell level through RNA FISH coupled with high-throughput imaging. Our enhancer tree models enabled the annotation of common germline risk variants associated with pancreas diseases, linking them to putative target genes in specific cell types. For pancreatic ductal adenocarcinoma, we found a stronger enrichment of disease susceptibility variants within acinar cell regulatory elements, despite ductal cells historically being assumed as the primary cell-of-origin. Our integrative approach--combining cell type-specific enhancer-promoter interaction mapping, computational models, and single-cell enhancer perturbation assays--produced a robust resource for studying the genetic basis of pancreas disorders.

genomics↗

bioRxiv2024

Elongation factor P controls ribosomal frameshift of a Salmonella antimicrobial resistance gene

Ribosomes translate mRNAs by matching every 3-nucleotide sequence in mRNA, producing the corresponding proteins. As the amino acid sequence directly dictates the activity of the protein, frameshifts often lead to unexpected effects. Here, ribosome profiling reveals that the intracellular pathogen Salmonella Typhimurium suppresses frameshift in the ugtL antimicrobial resistance gene during translation. This suppression of frameshift is mediated by a ribosome pause occurring in a newly-identified overlapping gene, serving as a non-slip bump. Given that the pause site contains a poly-proline motif and can be resolved by elongation factor P, the removal of the ribosome pause by substituting the motif induces ribosome slippage in ugtL, resulting in UgtL frameshifted protein production. This renders Salmonella sensitive to antimicrobial peptides but, in turn, protects the MgtC virulence factor from the FtsH-mediated proteolysis, indicating that elongation factor P-dependent ribosome pause is required for controlling both full antimicrobial resistance and mouse virulence. These findings reveal a new regulatory mechanism in which ribosome pause controls the production of two different protein isoforms by suppressing ribosome slippage-mediated frameshift.

microbiology↗

bioRxiv2024

A HigB-like toxin promotes non-replicating Salmonella inside macrophages by inhibiting ribonuclease III

Many bacteria are often resistant to antibiotic treatment because they can slow down their growth rate, thereby attenuating the drugs effectiveness. A similar growth-rate control is observed in pathogens that infect and persist inside their hosts. The bacterial toxin-antitoxin (TA) system serves as a non-heritable phenotypic switch by slowing down growth through the expression of a toxin component. Here, we investigated a HigB-like type II toxin from the intracellular bacterial pathogen Salmonella Typhimurium. Unlike other HigB-like toxins that cleave ribosome-bound mRNAs, it does not exhibit endoribonuclease activity. Instead, it inhibits ribonuclease III, which mediates the initial cleavage for rRNA processing, by directly binding to the dsRNA-binding domain of RNase III, thereby decreasing ribosome assembly and bacterial growth. Given that the formation of HigB-like toxin-mediated non-replicating Salmonella within macrophages is RNase III-dependent, persister formation by inhibiting RNase III is a newly identified strategy for pathogens to survive within host cells.

microbiology↗

bioRxiv2024

The transcriptomic architecture of the human cerebral cortex

For over a century, scientists have been attempting to map the human cerebral cortex, however, they have not taken into account the complex molecular structure of the cortex, which is only beginning to be understood. Here, we parcellate the human cerebral cortex using a machine learning (ML) approach to define its transcriptomic architecture, revealing a multi-resolution organization across individuals. The transcriptomically-derived spatial patterns of gene expression separate the cortex into three major regions, frontal, temporal and parietooccipital, with smaller subregions appearing at lower levels of the transcriptomic hierarchy. The core regions, which remain stable across different hierarchical levels, are physiologically associated with language, emotion regulation, social cognition, motor and visuospatial processing and planning. Importantly, some core regions cross structural and anatomical boundaries identified in previous parcellations of the cortex, revealing that the transcriptomic architecture of the cortex is closely linked to human-specific higher cognitive function.

neuroscience↗

bioRxiv2023

Unraveling the mechanism of centromere inactivation in human aging

Aging involves a range of genetic, epigenetic, and physiological alterations. A key characteristic of aged cells is the loss of global heterochromatin, accompanied by a reduction in canonical histone levels. In this study, we track the fate of centromeres during aging in human cells. Our findings reveal that the centromeric histone H3 variant CENP-A is downregulated in aged cells, in a p53-dependent manner. We observe repression of centromeric noncoding transcription through an epigenetic mechanism via recruitment of a lysine-specific demethylase 1 (LSD1/KDM1A) to centromeres. This suppression results in defective de novo CENP-A loading at aging centromeres. By dual inhibition of p53 and LSD1/KDM1A in aged cells, we mitigate the reduction in centromeric proteins and centromeric transcripts, leading to mitotic rejuvenation of these cells. These results offer insights into a novel mechanism for centromeric inactivation during aging and provide potential strategies to reactivate centromeres.

molecular biology↗

bioRxiv2023

Input-specific bi-directional regulation of CA3 pyramidal cell excitability: implication in sequential memory formation

Neuronal excitability is a key determinant for recruitment of a neuron to an ensemble. High-frequency mossy fiber (MF) inputs induce a prolonged increase in the excitability of a CA3 pyramidal cell (called long-term potentiation of intrinsic excitability, LTP-IE), thereby weak perforant pathway (PP) inputs can induce long-term potentiation at PP synapses (PP-LTP). However, sustained hyperexcitability is detrimental, and a mechanism to reverse this primed state is necessary. Here, we show that burst firings of CA3 pyramidal cells elicited by PP or recurrent synaptic inputs reverse the MF-induced LTP-IE. Moreover, the high-frequency PP inputs to MF-primed CA3 pyramidal cells induced not only PP-LTP but also restored the high excitability state. Labeling a neuronal ensemble using c-fos promoter in animals exposed to a novel context, we found most CA3 ensemble cells exhibited increased excitability, indicative of LTP-IE. Moreover, when the animals experienced novel contexts twice with an interval, a substantial subset of putative twice-activated CA3 ensemble cells exhibited reduced excitability, implying depotentiation of LTP-IE. We developed an in silico model based on these experimental results and found that MF-induced LTP-IE and its depotentiation are critical for association of orthogonal neuronal ensembles representing temporally discontiguous events. HighlightsO_LIIt is unknown how non-overlapping ensembles are linked in the hippocampal CA3 area. C_LIO_LIMossy fiber inputs prime CA3 pyramidal cells by enhancing dendritic excitability. C_LIO_LIPerforant pathway (PP) inputs to the primed cells induce synaptic strengthening. C_LIO_LIAt the same time, the high excitability state is restored by PP inputs. C_LIO_LIThis learning rule may play a key role in sequence learning in CA3 network. C_LI

neuroscience↗

bioRxiv2023

Discrimination of finger movements by magnetomyography with optically pumped magnetometers

Optically pumped magnetometers (OPM) are quantum sensors that offer new possibilities to measure biomagnetic signals. In magnetomyography (MMG), compared to the current standard surface electromyography (EMG), OPM sensors offer the advantage of contactless measurements of muscle activity. However, little is known about the relative performance of OPM-MMG and EMG, e.g. in their ability to detect and classify finger movements. To address this, we recorded simultaneous OPM-MMG and EMG of finger flexor muscles for the discrimination of individual finger movements. Using a deep learning model for movement classification, we found that both sensor modalities were able to discriminate finger movements with above 89% accuracy. Furthermore, model predictions for the two sensor modalities showed high agreement in movement detection (85% agreement; Cohens kappa: 0.45). Our findings show that OPM sensors can be employed for reliable, contactless discrimination of finger movements and incentivize future applications of OPM in magnetomyography.

neuroscience↗

bioRxiv2023

Integrative single-cell characterization of frugivory adaptations in the bat kidney and pancreas

Frugivory evolved multiple times in mammals, including bats. However, the cellular and molecular components driving it remain largely unknown. Here, we used integrative single-cell sequencing on insectivorous and frugivorous bat kidneys and pancreases and identified key cell population, gene expression and regulatory element differences associated with frugivorous adaptation that also relate to human disease, particularly diabetes. We found an increase in collecting duct cells and differentially active genes and regulatory elements involved in fluid and electrolyte balance in the frugivore kidney. In the frugivorous pancreas, we observed an increase in endocrine and a decrease in exocrine cells and differences in genes and regulatory elements involved in insulin regulation. Combined, our work provides novel insights into frugivorous adaptation that also could be leveraged for therapeutic purposes.

genomics↗