Biology subjects

Ha, S.

Publications and source records attributed to Ha, S..

At least 19 recordsLinked to original sources

bioRxiv2026

GeneReL: A Large Language Model-Powered Platform for Gene Regulatory Relationship Extraction with Community Curation

MotivationGene regulatory networks provide fundamental insights into plant biology, yet extracting structured interaction data from scientific literature remains a significant bottleneck. Traditional manual curation cannot scale to meet the demands of modern research, while automated text mining approaches struggle with the complexity of gene nomenclature and relationship classification. Large language models offer promising capabilities for information extraction, but integrated platforms combining LLM extraction with community validation for plant regulatory databases remain scarce. ResultsWe developed GeneReL, an integrated platform combining LLM-based extraction with community-driven curation for gene regulatory networks in Arabidopsis thaliana. The system employs a tiered pipeline using Claude Haiku 4.5 for screening, Claude Sonnet 4 for extraction, and Claude Opus 4 for verification, along with a novel five-step gene normalization pipeline incorporating paper-text search and LLM-based disambiguation with UniProt annotations. The database contains 13,710 curated interactions across 51 relationship types, with 90.2% classified as high confidence based on linguistic certainty markers in source text. Comparison with IntAct reveals 86.8% of interactions are unique to our literature-derived database, demonstrating complementary coverage to existing resources. The web platform provides card-based browsing with voting capabilities, interactive network visualization using Cytoscape.js with locus-ID-based node consolidation, and administrative interfaces for curator review of ambiguous gene mappings. Availability and ImplementationGeneReL is freely accessible at https://generel.newgenes.me. Contactkangyangjae@gnu.ac.kr

bioinformatics↗

bioRxiv2026

Reactive Astrocytes Drive Extracellular Acidification to Mediate α-Synuclein Neurodegeneration

Astrocytes are increasingly recognized as key players in neurodegeneration1-3, yet the molecular mechanisms by which they drive disease remain elusive. Here, we uncover a fundamental pathway in which reactive astrocytes fuel neurodegeneration in -synucleinopathies--including Dementia with Lewy bodies and Parkinsons disease dementia--by acidifying the brains extracellular environment. We demonstrate that both human patient tissue and a gut-to-brain -synuclein mouse model exhibit accumulation of reactive astrocytes and extracellular acidosis. Mechanistically, we show that astrocytic lysosomal exocytosis releases acidic contents, driving a drop in pH that activates neuronal acid-sensing ion channel 1a (ASIC1a), resulting in neuronal loss and behavioral decline. Blocking this pathway--either by inhibiting astrocytic lysosomal exocytosis or genetically or pharmacologically targeting neuronal ASIC1a--mitigates pathology and rescues neurodegenerative phenotypes in vivo. These findings provide a conceptual advance by establishing a mechanistic link between glial inflammation, acid-base homeostasis, and neuronal vulnerability, and suggest that targeting astrocyte-driven acidification or ASIC1a signaling could offer new avenues for disease modification in -synucleinopathies.

neuroscience↗

bioRxiv2026

Microglial Lag3 Drives α-Synuclein--induced Neurotoxic Activated (A1) Astrocytes and Neurodegeneration

BackgroundNeuroinflammation and pathologic -synuclein (-syn) aggregation cooperate to drive dopaminergic neurodegeneration in Parkinsons disease, but the glial receptors that couple extracellular -syn to inflammatory cascades remain incompletely defined. Microglia express higher levels of lymphocyte activation gene 3 (Lag3) than neurons, yet the contribution of microglial Lag3 to -syn recognition, glial crosstalk, and neurodegeneration is unknown. MethodsBiochemical binding assays, live-cell imaging, cytokine profiling, and neuron-microglia-astrocyte co-culture paradigms were used to define Lag3-dependent -syn preformed fibril (PFF) binding, uptake, and microglial activation. To interrogate in vivo function, microglia-specific Lag3 conditional knockout mice (Lag3L/L-Cx3cr1CreER) and littermate controls received unilateral intrastriatal -syn PFF injections, followed by histological, biochemical, and behavioral assessments of -syn pathology, gliosis, nigrostriatal integrity, and motor performance. Results-syn PFFs bound microglial Lag3 with high specificity and nanomolar affinity and required Lag3 for efficient fibril internalization and induction of proinflammatory cytokines. Microglial Lag3 deficiency markedly blunted -syn PFF-evoked microglial activation, prevented cytokine-driven conversion of astrocytes into neurotoxic reactive A1 astrocytes, and abolished astrocyte-dependent neuronal death in vitro. In vivo, microglia-specific Lag3 deletion reduced cortical, striatal, and substantia nigra pS129 -syn pathology, suppressed microgliosis and A1 astrocyte induction, preserved substantia nigra dopaminergic neurons and striatal dopamine transporter/tyrosine hydroxylase expression, and ameliorated -syn PFF-induced motor deficits. ConclusionsThis study identifies microglial Lag3 as a key receptor linking extracellular -syn PFF recognition to inflammatory amplification, neurotoxic reactive A1 astrocyte conversion, and dopaminergic neurodegeneration. Together with prior work on neuronal Lag3, these findings support a cell-type-specific dual-axis model in which neuronal Lag3 mediates -syn propagation while microglial Lag3 drives glia-dependent neurotoxicity, positioning Lag3 as a promising precision therapeutic target in -synucleinopathies.

neuroscience↗

bioRxiv2025

Neuronal PARIS-STAT3 axis drives tau pathology and glial activation in Alzheimers disease

PARIS, a substrate of Parkin, accumulates in Parkinsons disease and promotes disease progression. Here, we demonstrate that PARIS also contributes to Alzheimers disease by elevating STAT3 transcriptional activity, thereby inducing tau pathology, hippocampal atrophy, and glial activation. Genetic depletion of Paris reduced tau phosphorylation and cognitive decline in tauopathy mice, whereas neuron-specific PARIS overexpression caused tau accumulation, gliosis, and memory impairment. In contrast, astrocyte-specific overexpression did not induce pathology, indicating that PARIS acts in neurons to drive tau phosphorylation and glial activation. The pathological features induced by neuronal PARIS overexpression were rescued by STAT3 inhibition, demonstrating that PARIS-STAT3 signaling underlies these effects. Moreover, Paris knockout did not alter pathology in the amyloid-driven mouse model, highlighting specificity for tau pathology. Together, these findings reveal PARIS as a neuronal regulator of STAT3 signaling that exacerbates tau-mediated neurodegeneration and identify the PARIS-STAT3 pathway as a potential therapeutic target in Alzheimers disease.

neuroscience↗

bioRxiv2025

Transcranial Direct Current Stimulation Modulates Resting Brain Hemodynamics and Autonomic Function: A Multimodal fNIRS-HRV Study

Transcranial direct current stimulation (tDCS) is a non-invasive neuromodulation technique that can influence brain activity and physiological function. We conducted a sham-controlled multimodal study to examine the effects of low-intensity (0.375 mA) bifrontal tDCS at rest in healthy adult males. The tDCS, with the anode placed over the left dorsolateral prefrontal cortex (DLPFC) and the cathode over the right DLPFC, was applied for twelve minutes, and outcomes were measured with functional near-infrared spectroscopy (fNIRS) for cortical hemodynamics and connectivity, heart rate variability (HRV), photoplethysmography (PPG) for autonomic function, and subjective surveys for emotional stress. The results showed that, compared to sham, active tDCS produced a decrease in oxyhemoglobin (HbO) concentration in the left DLPFC during stimulation, indicating reduced cortical oxygenation in the stimulated region. Functional connectivity analysis of fNIRS signals further revealed altered network connectivity, including modulation of intra- and inter-frontal connections, in the tDCS condition relative to sham. Concurrently, tDCS induced an increase in HRV indices such as RMSSD, SDNN, pNN50, and Poincare plot measures SD1 and SD2, reflecting enhanced parasympathetic activity and autonomic regulation. Participants in the active tDCS group also reported selective reductions in both state and trait anger after stimulation, whereas the sham group showed minimal changes in mood. These findings demonstrate that even at a subthreshold intensity, bifrontal tDCS at rest engaged neurovascular and neurovisceral mechanisms, coupling changes in prefrontal cortical activity with autonomic outflow and mood. This study provides new evidence of cortical-autonomic coupling during neuromodulation and suggests that low-intensity frontal tDCS may promote a calm physiological and emotional state. The results have implications for future research on brain stimulation in emotion regulation and for developing neuromodulation-based interventions to improve autonomic balance and mood in both healthy individuals and clinical populations.

neuroscience↗

bioRxiv2025

EZH1 Q571R-mediated chromatin compaction and its oncogenic potential in thyroid cancer

Dysregulation of Polycomb Repressive Complex 2 (PRC2) is established to contribute to cancer. Of its catalytic subunits, EZH1 and EZH2, EZH2 mutations in cancer have been extensively identified and studied, but the role of EZH1 in cancer remains largely unexplored. The recurrent presence of the EZH1Q571R mutation in follicular thyroid cancer suggests its involvement in tumor progression. Using EZH1 ChIP-seq, ATAC-seq, and H3K27me3 CUT&Tag, we demonstrated that EZH1Q571R enhances chromatin binding and compaction and stimulates PRC2-EZH1 catalytic activity, leading to increased H3K27me3 levels and repression of tumor suppressor genes. Purified PRC2-EZH1Q571R demonstrated a significant increase in histone methyltransferase activity compared to PRC2-EZH1WT via enhanced nucleosome binding and DNA compaction. Notably, this effect is particularly observed with EZH1Q571R but only to a lesser extent with the corresponding EZH2Q570R mutation, highlighting it as an EZH1-specific mechanism. We additionally demonstrated PRC2-EZH1Q571R to efficiently methylate H3K27 in pre-existing H3K36me2/3 nucleosomes, disrupting chromatin homeostasis. Our findings provide key insights into the molecular pathogenesis of EZH1Q571R-driven follicular thyroid cancer.

molecular biology↗

bioRxiv2025

Dissecting the molecular landscape of Parkinson's disease and Parkinson's disease dementia using highly efficient snRNA-seq (HIF-snRNA-seq)

This study presents a transcriptomic analysis of the cingulate cortex (CING) in Parkinsons disease (PD) and Parkinsons disease dementia (PDD) using a High-efficiency single-nucleus RNA sequencing (HiF-snRNA-seq) protocol optimized for post-mortem brain samples. RNA quality prediction, poly-A tailing, and dCas9-targeted depletion enabled analysis of 77 high-quality samples from 240 cases, yielding over 2 million nuclei classified into seven major cell types. Disease conditions revealed altered astrocyte and microglia proportions, implicating their roles in neuroinflammation. Differential expression analysis identified unique and shared genes across PD and PDD, linked to synaptic remodeling, stress responses, and inflammation. Stage-specific analysis uncovered tau-dependent early-stage genes and inflammation-associated late-stage genes. This study highlights the CINGs central role in PD and PDD pathophysiology, offering insights into disease mechanisms and identifying candidate genes and pathways for therapeutic and biomarker development.

genomics↗

bioRxiv2024

Wirelessly-Powered Ingestible Electronic Capsule for Non-invasive Gastrointestinal Optogenetics

Optogenetics enables the activation and inhibition of neurons with cell specificity. The gut harbors intricate networks of enteric and central neurons. Uncovering these neuronal pathways in vivo is challenging with traditional neuroscience probes due to the highly motile and harsh gut environment. Here we report the development of an ingestible electronic capsule for non-invasive optical gut stimulation (ICOPS) in rodents. ICOPS is powered wirelessly via a transmitter coil, dosed via oral gavage, and safely excreted without causing obstruction. ICOPS permits modular interchangeability of onboard light-emitting diodes (LEDs) for illumination. We exemplify this with optical irradiance at 470 nm, a commonly-used wavelength in optogenetics for activating channelrhodopsin2. ICOPS features a micro-LED ({micro}LED), a 460-turn coil wound around a ferrite core, and a resonating capacitor. We optimized the transmitting and receiving circuits to achieve maximum power transfer at low operating frequencies (45-140 kHz), overcoming challenges like loose coupling and misalignment. The capsule operates effectively at a distance up to 12 cm longitudinally, 9 cm laterally, and 75{degrees} rotational angle relative to the magnetic field. Specific absorption rate (SAR) calculations indicate transmitter-induced SAR levels within safe limits for the occupational environment at 6 Arms and 45 and 63 kHz frequencies ICOPS is robust and transits through the rat gastrointestinal (GI) tract in under 20 hours intact. We demonstrate in vivo functionality and viability of ICOPS using IVIS micro-computed tomography ({micro}CT). ICOPS could pave the way for non-invasive optogenetic interfacing of enteric neural circuits towards their use to regulate motility, visceral pain, and other gastrointestinal disorders.

bioengineering↗

bioRxiv2024

High-throughput gene expression analysis with TempO-LINC sensitively resolves complex brain, lung and kidney heterogeneity at single-cell resolution

We report the development and performance of a novel genomics platform, TempO-LINC, for conducting high-throughput transcriptomic analysis on single cells and nuclei. TempO-LINC works by adding cell-identifying molecular barcodes onto highly selective and high-sensitivity gene expression probes within fixed cells, without having to first generate cDNA. Using an instrument-free combinatorial-indexing approach, all probes within the same fixed cell receive an identical barcode, enabling the reconstruction of single-cell gene expression profiles across as few as several hundred cells and up to 100,000+ cells per run. The TempO-LINC approach is easily scalable based on the number of barcodes and rounds of barcoding performed; however, for the experiments reported in this study, the assay utilized over 5.3 million unique barcodes. TempO-LINC has a robust protocol for fixing and banking cells and displays high-sensitivity gene detection from multiple diverse sample types. We show that TempO-LINC has an observed multiplet rate of less than 1.1% and a cell capture rate of [~]50%. Although the assay can accurately profile the whole transcriptome (19,683 human or 21,400 mouse genes), it can be targeted to measure only actionable/informative genes and molecular pathways of interest - thereby reducing sequencing requirements. In this study, we applied TempO-LINC to profile the transcriptomes of 89,722 cells across multiple sample types, including nuclei from mouse lung, kidney and brain tissues. The data demonstrated the ability to identify and annotate at least 50 unique cell populations and positively correlate expression of cell type-specific molecular markers within them. TempO-LINC is a robust new single-cell technology that is ideal for large-scale applications/studies across thousands of samples with high data quality.

genomics↗

bioRxiv2024

FtsN coordinates septal peptidoglycan synthesis and degradation through self-interaction of SPOR in E. coli

The E. coli cell division protein FtsN was proposed to coordinate septal peptidoglycan (sPG) synthesis and degradation to ensure robust cell wall constriction without lethal lesions. Although the precise mechanism remains unclear, previous work highlights the importance of two FtsN domains: the E domain, which interacts with and activates the sPG synthesis complex FtsWIQLB, and the SPOR domain, which binds to denuded glycan (dnG) strands, key intermediates in sPG degradation. Here, we used single-molecule tracking of FtsN and FtsW (a proxy for the sPG synthesis complex FtsWIQLB) to investigate how FtsN coordinates the two opposing processes. We observed dynamic behaviors indicating that FtsNs SPOR domain binds to dnGs cooperatively, which both sequesters the sPG synthesis complex on dnG (termed as the dnG-track) and protects dnGs from degradation by lytic transglycosylases (LTs). The release of the SPOR domain from dnGs leads to activating the sPG synthesis complex on the sPG-track and simultaneously exposing those same dnGs to degradation. Furthermore, FtsNs SPOR domain self-interacts and facilitates the formation of a multimeric sPG synthesis complex on both tracks. The cooperative self-interaction of the SPOR domain creates a sensitive switch to regulate the partitioning of FtsN between the dnG- and sPG-tracks, thereby controlling the balance between sequestered and active populations of the sPG synthesis complex. As such, FtsN coordinates sPG synthesis and degradation in space and time.

microbiology↗

bioRxiv2024

Generative Adversarial Implicit Successor Representation

We propose a novel method to implicitly encode Successor Representations (SRs) using a Generative Adversarial Network (GAN). SRs are a method to encode states of the environment in terms of their predictive relationships with other states, which can be used to predict long-term future rewards. In standard explicit methods, the value of SR is found from an explicit map between future states after an action or to find an approximate function. Instead, our method encodes SR implicitly using a GAN. The distribution of samples generated by the GAN system approximates the successor representation. We also suggest an action decision procedure for the implicit encoding of SR. The system makes the decision using an analysis-by-synthesis procedure that it attempts to synthesize a sample that can explain the action decision constraints of the current and target states. Our system is different from the classical SR in several points. It can sample actual samples reflecting SR distribution, which is not easy for explicit models. It can also get around the issue of explicitly representing probabilities or successor representation values and doing math over them. We tested our system in a toy environment, where the agent could learn the implicit successor representation successfully and use it for action decisions.

neuroscience↗

bioRxiv2023

An ACE2 decamer viral trap as a durable intervention solution for current and future SARS-CoV

The capacity of SARS-CoV-2 to evolve poses challenges to conventional prevention and treatment options such as vaccination and monoclonal antibodies, as they rely on viral receptor binding domain (RBD) sequences from previous strains. Additionally, animal CoVs, especially those of the SARS family, are now appreciated as a constant pandemic threat. We present here a new antiviral approach featuring inhalation delivery of a recombinant viral trap composed of ten copies of angiotensin-converting enzyme 2 (ACE2) fused to the IgM Fc. This ACE2 decamer viral trap is designed to inhibit SARS-CoV-2 entry function, regardless of viral RBD sequence variations as shown by its high neutralization potency against all known SARS-CoV-2 variants, including Omicron BQ.1, BQ.1.1, XBB.1 and XBB.1.5. In addition, it demonstrates potency against SARS-CoV-1, human NL63, as well as bat and pangolin CoVs. The multivalent trap is effective in both prophylactic and therapeutic settings since a single intranasal dosing confers protection in human ACE2 transgenic mice against viral challenges. Lastly, this molecule is stable at ambient temperature for more than twelve weeks and can sustain physical stress from aerosolization. These results demonstrate the potential of a decameric ACE2 viral trap as an inhalation solution for ACE2-dependent coronaviruses of current and future pandemic concerns.

microbiology↗

bioRxiv2023

Mature microRNA-binding protein QKI promotes microRNA-mediated gene silencing

Although Argonaute (AGO) proteins have been the focus of microRNA (miRNA) studies, we observed AGO-free mature miRNAs directly interacting with RNA-binding proteins, implying the sophisticated nature of fine-tuning gene regulation by miRNAs. To investigate microRNA-binding proteins (miRBPs) globally, we analyzed PAR-CLIP data sets to identify RBP quaking (QKI) as a novel miRBP for let-7b. Potential existence of AGO-free miRNAs were further verified in genetically engineered AGO-depleted human and mouse cells. We have shown that QKI serves as an auxiliary factor empowering AGO2/let-7b-mediated gene silencing. Depletion of QKI decreases interaction of AGO2 with let-7b and target mRNA, consequently controlling target mRNA decay. QKI, however, also suppresses the dissociation of let-7b from AGO2, and slows assembly of AGO2/miRNA/target mRNA complexes at the single-molecule level. We also revealed that QKI suppresses cMYC expression at post-transcriptional level, and decreases proliferation and migration of HeLa cells, demonstrating that QKI is a tumor suppressor gene by in part augmenting let-7b activity. Our data show that QKI is a new type of RBP implicated in the versatile regulation of miRNA-mediated gene silencing.

biochemistry↗

bioRxiv2023

Synthetic lipid rafts formed by cholesterol nano-patch induce T cell activation

Lipid rafts cluster at the immunological synapse during activation and serve as signaling hubs in T cells. While cholesterol plays a crucial role in lipid raft formation, the impact of their spatial configuration remains less understood. Here, we introduce programmable DNA origami, cholesterol nano-patch, for nanoscale spatial control of cholesterols on live T cell membranes, enabling us to elucidate their roles in lipid raft formation, receptor activation, and intracellular signaling. We demonstrate that CNPs with high-density cholesterol arrangements efficiently bind to the T cell plasma membrane and form a large and polarized coalescence, leading to the stabilization of ordered lipid membrane domains by increasing the local concentration of cholesterols. These synthetic lipid rafts colocalize with flotillin-1, a raft-associated protein, and promote the membrane reorganization and physical segregation of key signaling molecules, such as T cell receptors, Lck and LAT kinases, and CD45. Consequently, this leads to early T cell activation in the absence of antigenic stimulation. Our DNA origami approach demonstrates the direct role of cholesterol in lipid raft formation, leading to membrane phase separation and protein reorganization, which are sufficient to induce early T cell activation.

bioengineering↗

bioRxiv2022

Korea4K: whole genome sequences of 4,157 Koreans with 107 phenotypes derived from extensive health check-ups

We present 4,157 whole-genome sequences (Korea4K) coupled with 107 health check-up parameters as the largest whole genomic resource of Koreans. Korea4K provides 45,537,252 variants and encompasses most of the common and rare variants in Koreans. We identified 1,356 new geno-phenotype associations which were not found by the previous Korea1K dataset. Phenomics analyses revealed 24 genetic correlations, 1,131 pleiotropic variants, and 127 causal relationships from Mendelian randomization. Moreover, the Korea4K imputation reference panel showed a superior imputation performance to Korea1K. Collectively, Korea4K provides the most extensive genomic and phenomic data resources for discovering clinically relevant novel genome-phenome associations in Koreans.

genomics↗

bioRxiv2022

A general model-based causal inference overcomes the curse of synchrony and indirect effect

To identify causation, model-free inference methods, such as Granger Causality, have been widely used due to their flexibility. However, they have difficulty distinguishing synchrony and indirect effects from direct causation, leading to false predictions. To overcome this, model-based inference methods that test the reproducibility of data with a specific mechanistic model to infer causality were developed. However, they can only be applied to systems described by a specific model, greatly limiting their applicability. Here, we address this limitation by deriving an easily testable condition for a general monotonic ODE model to reproduce time-series data. We built a user-friendly computational package, GOBI (General ODE-Based Inference), which is applicable to nearly any monotonic system with positive and negative regulations described by ODE. GOBI successfully inferred positive and negative regulations in various networks at both the molecular and population levels, unlike existing model-free methods. Thus, this accurate and broadly applicable inference method is a powerful tool for understanding complex dynamical systems.

systems biology↗

bioRxiv2022

A TMT-based quantitative proteomics approach toward α-syn PFF associated Lewy Body Dementia (LBD) using α-syn PFF-injected mouse brain tissues

The aggregation of -synuclein in the nervous system leads to a class of neurodegenerative disorders termed -synucleinopathies. A form of primary degenerative dementia called Lewy body dementia (LBD) often develops in the case these aggregations develop into intracellular inclusions called Lewy bodies (LB) and Lewy neurites (LN). Despite the high frequency of LBD, being the leading cause of dementia following Alzheimers disease (AD), there is relatively little information discovered about its pathological pathway or diagnostic criteria. In this report, we attempt to address such shortcomings via utilizing a proteomic approach to identify the proteomic changes following intrastriatal injection of -synuclein preformed fibril (-syn PFF). Through mass spectrometry, we have identified a total of 179 proteins that were either up- or down-regulated at different time points, with the four proteins - TPP3, RAB10, CAMK2A, and DYNLL1 - displaying the most significant changes throughout the timeframe. Further examining the modulated proteins with network-based enrichment analyses, we have found that 1) the most significantly associated neurodegenerative pathways were Parkinsons (pV = 3.0e-16) and Huntingtons (pV = 1.9e-15) disease, and 2) the majority of molecular functions specific to the pathology only appeared at later time points. While these results do not expose a conclusive biomarker for LBD, they suggest a potential framework that may be utilized to diagnose and differentiate LBD pathology from other forms of dementia by focusing on the cortical proteomic changes which occur in a later time span.

cell biology↗

bioRxiv2022

Proteome-wide microarray-based screening of PAR-binding proteins

Poly(ADP-ribose) (PAR) plays a crucial role in intracellular signaling and scaffolding through covalent modification or non-covalent binding to target proteins. The non- covalent binding PARylome has not been extensively characterized. Here we performed a PAR-binding screen using a human protein microarray that covers most of the human proteome to characterize the non-covalent binding PARylome. A total of 356 PAR- binding proteins were identified. The PAR-binding PARylome suggests that PAR- binding regulates a variety of biological processes beyond well-characterized DNA damage signaling and DNA repair. Proteins that may be reprogrammed by PAR-binding include signaling molecules, transcription factors, nucleic acid binding proteins, calcium binding proteins, ligases, oxidoreductases, enzymes, transferases, hydrolases, and receptors. The global database of PAR-binding proteins that we established will be a valuable tool for further in-depth analysis of the role of PARylation in a wide range of biological contexts.

biochemistry↗