bioRxiv Science⌕ Search

Biology subjects

Yoon, K.-J.

Publications and source records attributed to Yoon, K.-J..

11 recordsLinked to original sources
bioRxiv2026

Deep Learning-Guided Holotomography Reveals Early Structural Remodelling During Pluripotency Exit

Real-time assessment of human pluripotent stem cell (hPSC) quality is critical for reproducibility and safety in regenerative medicine, yet current methods are invasive, labor-intensive, or highly operator-dependent. We present DeepHOPE (Deep-learning-guided Holotomography for Pluripotency Evaluation), a non-invasive, automatizable, and statistics-driven platform that integrates three-dimensional (3D) refractive index imaging with deep learning to assess pluripotency. DeepHOPE performs robustly across diverse contexts, including germ-layer differentiation, retinoic acid-induced differentiation, and mid-reprogramming cultures, enabling streamlined cell production workflows and improving the efficiency of midbrain dopaminergic neuron differentiation through informed colony selection. Mechanistically, DeepHOPE detects minute colony-scale topological changes that precede molecular loss of pluripotency. These early changes are associated with rapid F-actin remodeling, including apical-to-basal redistribution during early differentiation. Consistent with a functional role for cytoskeletal regulation in state transitions, sustained reduction of actomyosin tension decreases pluripotency, identifying cytoskeletal dynamics as an upstream determinant of early pluripotency exit. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=115 SRC="FIGDIR/small/720508v1_ufig1.gif" ALT="Figure 1"> View larger version (45K): org.highwire.dtl.DTLVardef@1d73508org.highwire.dtl.DTLVardef@1dab29aorg.highwire.dtl.DTLVardef@1039f76org.highwire.dtl.DTLVardef@da29a1_HPS_FORMAT_FIGEXP M_FIG C_FIG

developmental biology↗
bioRxiv2026

Deep Plasma Proteomics Coupled with Functional Genomics Reveals Drivers of Parkinson's Disease Progression and Levodopa Response

Parkinsons disease (PD) is a progressive neurodegenerative disorder lacking disease-modifying therapies, and its clinical management is limited by the absence of accessible biomarkers for tracking disease progression and treatment response. To map these complex disease trajectories, we implemented an ultra-deep plasma proteomics workflow integrating Mag-Net extracellular vesicle enrichment with Orbitrap Astral mass spectrometry to profile longitudinal samples from PD patients. This approach quantified 6,481 plasma proteins at unprecedented depth in PD studies, revealing distinct signatures directly associated with disease duration and dopaminergic therapy exposure. Candidate biomarkers were subsequently validated in an independent cohort using ELISA, demonstrating robust predictive utility in AI-driven prediction models. To uncover the mechanistic drivers underlying these systemic changes, we intersected our proteomic data with novel proteome-wide gene overexpression perturbation screens designed to identify regulators of alpha-synuclein pre-formed fibril (PFF) uptake and PFF-induced neuronal toxicity. Finally, an integrative network analysis combining three independent proteome-wide assays revealed that key pathological hubs, such as CD14, IFNG, and PLAT, are targets of currently approved pharmacological agents. Collectively, these findings provide a comprehensive, systems-level map for PD biomarker discovery and highlight druggable pathways to advance precision medicine strategies.

neuroscience↗
bioRxiv2025

A machine learning framework for predicting and modulating condition-dependent protein phase separation

Protein phase separation is a fundamental process in organizing membraneless organelles and is implicated in a wide range of pathological conditions. Importantly, rather than being a static feature of specific proteins, phase separation is a condition-dependent phenomenon governed by environmental parameters, including protein concentration, temperature, and solvent composition. However, most existing machine learning models infer phase-separation propensity solely from amino-acid sequences, failing to capture these context-dependent behaviors. Here, we present LLPSense, a machine learning framework that integrates pre-trained protein language model embeddings with environmental parameters to achieve accurate, condition-aware predictions of protein phase separation. We demonstrate LLPSenses predictive power and utility through three key experimental demonstrations. First, the model revealed that SGTA, previously unrecognized as a phase-separating protein, exhibits complex, temperature-dependent reentrant phase behavior. Second, LLPSense accurately predicted mutations in -synuclein that either enhance or suppress phase separation, enabling systematic mapping of residues potentially relevant to Parkinsons disease. Third, using model-guided mutagenesis, we inverted the phase behavior of UBQLN4, shifting it from high-temperature to low-temperature separation. Collectively, LLPSense provides a robust computational tool for interrogating the condition-dependent landscape of protein phase separation, enabling mechanistic studies of disease-associated phase separation and the rational design of programmable condensates.

bioinformatics↗
bioRxiv2025

Nano-Sieve Enabled On Chip Concentration and Multiplexed Detection of Raman-Dye Labeled Nanocubes

To address persistent challenges in molecular detection--particularly signal variability, low analyte abundance, and poor spectral reproducibility--a portable bead-stacked nano-sieve platform was engineered to enable robust on-chip Raman signal amplification using silver nanocubes (AgNCs) functionalized with Raman-active thiolated ligands. The microdevice integrates deformable elastomeric channels with magnetic bead stacking to immobilize and concentrate AgNCs within confined nanoscopic junctions, thereby generating densely packed plasmonic hot-spot architectures that markedly enhance Surface-Enhanced Raman Scattering (SERS) efficiency. Time-resolved measurements revealed substantial increases in Raman intensity for both 4-mercaptobenzoic acid (4-MBA) and 4-aminothiophenol (4-ATP), outperforming corresponding off-chip assays and maintaining stable enhancement factors over a four-hour operational window. Furthermore, the platform successfully demonstrated multiplexed sensing capability: mixed-analyte experiments containing 4-MBA, 4-ATP, and cysteamine produced clearly distinguishable spectral fingerprints, even under serial dilution and dynamically pulsed sample delivery. Collectively, these results establish the bead-stacked nano-sieve as a versatile, label-free, and scalable diagnostic architecture capable of achieving high sensitivity and spectral fidelity under continuous-flow conditions. Our technology holds significant promise for real-time, non-invasive molecular surveillance, point-of-care diagnostics, and early disease monitoring in resource-limited or field-deployable settings.

bioengineering↗
bioRxiv2025

Organelle-Aware Representation Learning Enables Label-Free Detection of Mitochondrial Dysfunction in Live Human Neurons

Mitochondrial dysfunction is a convergent hallmark of neurodegenerative diseases and represents a promising biomarker for early diagnosis and therapy. However, current in vitro assays rely on fluorescence or electron microscopy, which are invasive, low-throughput, and incompatible with longitudinal analysis. Here, we present a noninvasive framework by integrating label-free optical diffraction tomography (ODT) with organelle-aware representation learning to detect subtle mitochondrial dysfunction in live human induced pluripotent stem cell (hiPSC)-derived neurons. Through virtual staining of the nuclei, lysosomes, and mitochondria, we establish two complementary and interpretable classification pipelines: a deep learning model with organelle-aware encoder and a logistic regression model on morphometric descriptors. Both models achieve approximately 85% accuracy: the deep model provides end-to-end prediction with minimal feature engineering, whereas the logistic regression model offers a more interpretable, feature-based approach. To our knowledge, this is the first demonstration of ODT-based organelle-resolved virtual staining in live human neurons, establishing a scalable, non-invasive platform for mitochondrial disease modeling, drug discovery, and neurodegeneration research.

neuroscience↗
bioRxiv2023

Identification and Analysis of SARS-CoV-2 Mutation and Subtype using 2x tiled Primer Set with Oxford Nanopore Technologies Sequencing for Enhanced Variant Detection and Surveillance in Seoul, Korea

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a respiratory virus that contains RNA as its genetic material and has caused a global pandemic since its outbreak in 2020. This virus has many mutations, some of which can reduce the effectiveness of existing vaccines. Therefore, next-generation sequencing (NGS) is necessary to accurately identify new mutations. Current NGS analysis of SARS-CoV-2 uses the amplicon analysis method through a multiplex polymerase chain reaction. This study collected and validated RNA samples from patients who tested positive for SARS-CoV-2 from April to July 2022, and selected 613 samples for sequencing. The findings demonstrate the importance of long-read-based NGS analysis and 2x tiled primer set for identifying full SARS-CoV-2 genome sequence with new mutations and understanding the correlation between viral genotypes and patient characteristics for the effective management of SARS-CoV-2.

microbiology↗
bioRxiv2023

Adolescent neurostimulation of dopamine circuit reverses genetic deficits in frontal cortex function

Dopamine system dysfunction is commonly implicated in adolescent-onset neuropsychiatric disorders. Although psychosis symptoms can be alleviated by antipsychotics, cognitive symptoms remain unresponsive to such pharmacological treatments and novel research paradigms investigating the circuit substrates underlying cognitive deficits are critically needed. The frontal cortex and its dopaminergic input from the midbrain are implicated in cognitive functions and undergo maturational changes during adolescence. Here, we used mice carrying mutations in the Arc or DISC1 genes to model mesofrontal dopamine circuit deficiencies and test circuit-based neurostimulation strategies to restore cognitive functions. We found that in a memory-guided spatial navigation task, frontal cortical neurons were activated coordinately at the decision-making point in wild-type but not Arc mutant mice. Chemogenetic stimulation of midbrain dopamine neurons or optogenetic stimulation of frontal cortical dopamine axons in a limited adolescent period consistently reversed genetic defects in mesofrontal innervation, task-coordinated neuronal activity, and memory-guided decision-making at adulthood. Furthermore, adolescent stimulation of dopamine neurons also reversed the same cognitive deficits in DISC1 mutant mice. Our findings reveal common mesofrontal circuit alterations underlying the cognitive deficits caused by two different genes and demonstrate the feasibility of adolescent neurostimulation to reverse these circuit and behavioral deficits. These results may suggest developmental windows and circuit targets for treating cognitive deficits in neurodevelopmental disorders.

neuroscience↗
bioRxiv2020

An Integrated Systems Biology Approach Identifies the Proteasome as a Critical Host Machinery for ZIKV and DENV Replication

The Zika (ZIKV) and dengue (DENV) flaviviruses exhibit similar replicative processes but distinct clinical outcomes. A systematic understanding of virus-host protein-protein interaction networks can reveal cellular pathways critical to viral replication and disease pathogenesis. Here we employed three independent systems biology approaches toward this goal. First, protein array analysis of direct interactions between individual ZIKV/DENV viral proteins and 20,240 human proteins revealed multiple conserved cellular pathways and protein complexes, including proteasome complexes. Second, an RNAi screen of 10,415 druggable genes to identify host proteins required for ZIKV infection uncovered proteasome proteins. Third, a high-throughput screening of 6,016 bioactive compounds for ZIKV inhibitors yielded 134 effective compounds, including six proteasome inhibitors that suppress both ZIKV and DENV replication. Integrative analyses of these orthogonal datasets pinpoints proteasome as critical host machinery for ZIKV/DENV replication. Our study provides multi-omics datasets for further studies of flavivirus-host interactions, disease pathogenesis, and new drug targets.

microbiology↗
bioRxiv2020

Structure and immune recognition of the porcine epidemic diarrhea virus spike protein

Porcine epidemic diarrhea virus is an alphacoronavirus responsible for significant morbidity and mortality in pigs. A key determinant of viral tropism and entry, the PEDV spike protein is a key target for the host antibody response and a good candidate for a protein-based vaccine immunogen. We used electron microscopy to evaluate the PEDV spike structure, as well as pig polyclonal antibody responses to viral infection. The structure of the PEDV spike reveals a configuration similar to that of HuCoV-NL63. Several PEDV protein-protein interfaces are mediated by non-protein components including a glycan at Asn264 and two bound palmitoleic acid molecules. The polyclonal antibody response to PEDV infection shows a dominance of epitopes in the S1 region. This structural and immune characterization provides new insights into coronavirus spike stability determinants and explores the immune landscape of viral spike proteins.

biophysics↗
bioRxiv2019

Persistent Cyfip1 expression is required to maintain the adult subventricular zone neurogenic niche

Neural stem cells (NSCs) persist throughout life in the subventricular zone (SVZ) niche of the lateral ventricles as B1 cells. Maintaining this population of NSCs depends on the balance between quiescence and self-renewing or self-depleting proliferation. Interactions between B1 cells and the surrounding niche are important in regulating this balance, but the mechanisms governing these processes have not been fully elucidated in adult mammals. The cytoplasmic FMRP-interacting protein (CYFIP1) regulates apical-basal polarity in the embryonic brain. Loss of Cyfip1 during embryonic development in mice disrupts the embryonic niche and affects cortical neurogenesis. However, a direct role for Cyfip1 in the regulation of adult NSCs has not been established. Here, we demonstrate that Cyfip1 expression is preferentially localized to B1 cells in the adult SVZ. Loss of Cyfip1 in the embryonic mouse brain results in altered adult SVZ architecture and expansion of the adult B1 cell population at the ventricular surface. Furthermore, acute deletion of Cyfip1 in adult NSCs results in a rapid change in adherens junction proteins as well as increased proliferation and the number of B1 cells at the ventricular surface. Together, these data indicate that CYFIP1 plays a critical role in the formation and maintenance of the adult SVZ niche and, furthermore, deletion of Cyfip1 unleashes the capacity of adult B1 cells for symmetric renewal to increase the adult NSC pool.\n\nSIGNIFICANCENeural stem cells (NSCs) persist in the subventricular zone (SVZ) of the lateral ventricles in adult mammals and their population is determined by the balance between quiescence and self-depleting or renewing cell division. The mechanisms regulating their biology are not fully understood. This study establishes that the cytoplasmic FMRP interacting protein 1 (Cyfip1) regulates NSC fate decisions in the adult SVZ and NSCs that are quiescent or typically undergo self-depleting divisions retain the ability to self-renew in the adult. This contributes to our understanding of how adult NSCs are regulated throughout life and has potential implications for human brain disorders.

developmental biology↗