Biology subjects

Park, Y.-G.

Publications and source records attributed to Park, Y.-G..

8 recordsLinked to original sources

bioRxiv2026

VARION: A Network Propagation Framework for Individual Patient Somatic Mutation Interpretation in Cancer Molecular Subtyping

Accurate molecular subtyping of individual cancer patients from somatic mutation data remains a challenge in precision oncology research. Existing network-based stratification (NBS) methods treat all mutations equivalently, require full-cohort batch processing, and do not demonstrate generalization to independent datasets without retraining. To address this, we present variant interpretation via the adaptive network pRopagatION (VARION), which integrates population-level variant constraint scoring with protein-protein interaction (PPI) network topology. The Adaptive Topology-aware Random Walk with Restart (ATR-RWR) algorithm weights each mutated gene by {varphi}g = {surd}(GIS(g) x {rho}topo(g)), where GIS (Gene Intolerance Score) reflects population-level functional constraint, propagated across a shared PPI network; subtype assignment then uses cosine similarity to TCGA-derived reference centroids, enabling real-time single-patient classification. Across ten TCGA cancer cohorts (n = 2,417), VARION achieved 77.7% accuracy for ovarian cancer (OV), 69.5% for glioblastoma (GBM), 90.2% for cholangiocarcinoma (CHOL), and 75.4% for gastric cancer (STAD). A controlled benchmark applying two alternative clustering methods (PyNBS; a dense autoencoder) to identical ATR-RWR propagation matrices recovered no significant driver enrichment (OR = 1.79 and 1.52, n.s.), versus OR = 144.29 (p = 1.77x10^-12) for VARION, confirming that the GIS-weighted centroid architecture, not propagation alone, drives performance; generalization without retraining was further confirmed in two independent cohorts (ICGC CCA, n = 396; PCAWG, n = 110; OR = {infty}, p < 5x10^-9). Together, these results indicate that VARION's GIS-weighted centroid architecture enables individual-patient molecular subtyping that outperforms existing NBS and graph-learning clustering approaches, with high sensitivity for clinically actionable rare subtypes and robust cross-platform generalization.

bioinformatics↗

bioRxiv2026

Whole-brain protein profiling using organ-scale multiplexed immunolabeling and image co-registration

Proteins are major drivers of biological functions. Single-cell, organ-scale multiplexed protein imaging can reveal high-dimensional molecular and structural features of individual cells and their interactions, enabling an in-depth understanding of complex biological systems. However, achieving such imaging has remained elusive due to hurdles in multiplexed immunolabeling (mIF) of intact organs and integrative image analysis. Here, we present 3D CYCLIC, an organ-scale multiplexed immunolabeling technique, and TACTIC, a single-cell-level, organ-scale image co-registration algorithm. 3D CYCLIC combines ultrafast, versatile 3D immunolabeling with a cleavable crosslinker that preserves signals by protecting bound antibodies during optical clearing while enabling their detachment for subsequent rounds of immunolabeling. TACTIC uses deep warping networks coupled with a propagation-based cell-pair search to co-register individual cells across whole-brain images acquired from multiple rounds of 3D CYCLIC labeling. 3D CYCLIC enabled 6-plex protein profiling of a mouse brain hemisphere, with images that can be combined with TACTIC for integrative analysis. 3D CYCLIC and TACTIC are versatile techniques and will therefore enable a holistic, unbiased understanding of diverse organs.

neuroscience↗

bioRxiv2026

3DBrainOne: an integrated end-to-end platform for 3D histological analysis of whole mouse brains

Three-dimensional (3D) whole-organ imaging and analysis at cellular resolution (termed 3D histology) provide profound insights into the organization and interactions of cells throughout organs. However, the quantitative analysis of these massive datasets remains a significant bottleneck due to the lack of integrated, user-friendly tools. Here, we present 3DBrainOne, an end-to-end ImageJ plugin that streamlines the essential 3D histological analysis of the mouse brain--from raw image preprocessing to region-wise quantification--within a single platform. 3DBrainOne features a robust whole-brain cell-counting module that uses a Difference-of-Gaussians (DoG) blob detection algorithm followed by a ResNet18-based deep learning classifier, enabling high-fidelity automatic whole-brain cell counting with a graphical user interface (GUI) for visual inspection and manual curation of analysis results. 3DBrainOne also supports multi-channel colocalization analysis. Furthermore, this platform includes modules for atlas alignment and brain-region-wise volumetric quantification, enabling brain region-resolved cell counting and structural analyses. As an ImageJ plugin, 3DBrainOne is compatible with a range of operating systems and hardware. In summary, 3DBrainOne is an integrated, versatile, and easy-to-use platform that will facilitate 3D histological analyses in experimental neuroscience.

neuroscience↗

bioRxiv2020

Divergent pallidal pathways underlying distinct Parkinsonian behavioral deficits

The basal ganglia are a group of subcortical nuclei that regulates motor and cognitive functions1,2. Recent identification of neuronal heterogeneity in the basal ganglia suggests that functionally distinct neural circuits defined by their efferent projections exist even within the same nuclei3-5. This distinction may account for a multitude of symptoms associated with basal ganglia disorders such as Parkinsons disease (PD)6,7. However, our incomplete understanding of the basal ganglia functional organization has hindered further investigation of individual circuits that may underlie different behavioral symptoms in disease states. Here we functionally define two distinct classes of parvalbumin-expressing neurons in the mouse external globus pallidus (GPe-PV) embedded within discrete neural pathways and establish their contributions to different Parkinsonian behavioral deficits. We find that GPe-PV neurons projecting to the substantia nigra pars reticulata (SNr) or parafascicular thalamus (PF) undergo different electrophysiological adaptations in response to dopamine depletion. Furthermore, counteracting these adaptations in each population can selectively alleviate movement deficits or behavioral inflexibility in a Parkinsonian mouse model. Our findings provide a novel framework to understand the circuit basis of separate behavioral symptoms in Parkinsonian state which could provide better strategies for the treatment of PD.

neuroscience↗

bioRxiv2020

Cellular Anatomy of the Mouse Primary Motor Cortex

An essential step toward understanding brain function is to establish a cellular-resolution structural framework upon which multi-scale and multi-modal information spanning molecules, cells, circuits and systems can be integrated and interpreted. Here, through a collaborative effort from the Brain Initiative Cell Census Network (BICCN), we derive a comprehensive cell type-based description of one brain structure - the primary motor cortex upper limb area (MOp-ul) of the mouse. Applying state-of-the-art labeling, imaging, computational, and neuroinformatics tools, we delineated the MOp-ul within the Mouse Brain 3D Common Coordinate Framework (CCF). We defined over two dozen MOp-ul projection neuron (PN) types by their anterograde targets; the spatial distribution of their somata defines 11 cortical sublayers, a significant refinement of the classic notion of cortical laminar organization. We further combine multiple complementary tracing methods (classic tract tracing, cell type-based anterograde, retrograde, and transsynaptic viral tracing, high-throughput BARseq, and complete single cell reconstruction) to systematically chart cell type-based MOp input-output streams. As PNs link distant brain regions at synapses as well as host cellular gene expression, our construction of a PN type resolution MOp-ul wiring diagram will facilitate an integrated analysis of motor control circuitry across the molecular, cellular, and systems levels. This work further provides a roadmap towards a cellular resolution description of mammalian brain architecture.

neuroscience↗

bioRxiv2019

Brain-wide mapping of contextual fear memory engram ensembles supports the dispersed engram complex hypothesis

GRAPHICAL ABSTRACT\n\nO_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=145 SRC=\"FIGDIR/small/668483v1_ufig1.gif\" ALT=\"Figure 1\">\nView larger version (34K):\norg.highwire.dtl.DTLVardef@1957e90org.highwire.dtl.DTLVardef@1a83d79org.highwire.dtl.DTLVardef@5bd5d5org.highwire.dtl.DTLVardef@133c879_HPS_FORMAT_FIGEXP M_FIG C_FIG SUMMARYNeuronal ensembles that hold specific memory (memory engrams) have been identified in the hippocampus, amygdala, and cortex. It has been hypothesized that engrams for a specific memory are distributed among multiple brain regions that are functionally connected. Here, we report the hitherto most extensive engram map for contextual fear memory by characterizing activity-tagged neurons in 409 regions using SHIELD-based tissue phenotyping. The mapping was aided by a novel engram index, which identified cFos+ brain regions holding engrams with a high probability. Optogenetic manipulations confirmed previously known engrams and revealed new engrams. Many of these engram holding-regions were functionally connected to the CA1 or amygdala engrams. Simultaneous chemogenetic reactivation of multiple engrams, which mimics natural memory recall, conferred a greater level of memory recall than reactivation of a single engram ensemble. Overall, our study supports the hypothesis that a memory is stored in functionally connected engrams distributed across multiple brain regions.

neuroscience↗

bioRxiv2019

Ultrafast immunostaining of organ-scale tissues for scalable proteomic phenotyping

Studying the function and dysfunction of complex biological systems necessitates comprehensive understanding of individual cells. Advancements in three-dimensional (3D) tissue processing and imaging modalities have enabled rapid visualization and phenotyping of cells in their spatial context. However, system-wide interrogation of individual cells within large intact tissue remains challenging, low throughput, and error-prone owing to the lack of robust labeling technologies. Here we introduce a rapid, versatile, and scalable method, eFLASH, that enables complete and uniform labeling of organ-scale tissue within one day. eFLASH dynamically modulates chemical transport and reaction kinetics to establish system-wide uniform labeling conditions throughout the day-long labeling period. This unique approach enables the same protocol to be compatible with a wide range of tissue types and probes, enabling combinatorial molecular phenotyping across different organs and species. We applied eFLASH to generate quantitative maps of various cell types in mouse brains. We also demonstrated multidimensional cell profiling in a marmoset brain block. We envision that eFLASH will spur holistic phenotyping of emerging animal models and disease models to help assess their functions and dysfunctions.

bioengineering↗

bioRxiv2019

Scalable image processing techniques for quantitative analysis of volumetric biological images from light-sheet microscopy

Here we describe an image processing pipeline for quantitative analysis of terabyte-scale volumetric images of SHIELD-processed mouse brains imaged with light-sheet microscopy. The pipeline utilizes open-source packages for destriping, stitching, and atlas alignment that are optimized for parallel processing. The destriping step removes stripe artifacts, corrects uneven illumination, and offers over 100x speed improvements compared to previously reported algorithms. The stitching module builds upon Terastitcher to create a single volumetric image quickly from individual image stacks with parallel processing enabled by default. The atlas alignment module provides an interactive web-based interface that automatically calculates an initial alignment to a reference image which can be manually refined. The atlas alignment module also provides summary statistics of fluorescence for each brain region as well as region segmentations for visualization. The expected runtime of our pipeline on a whole mouse brain hemisphere is 1-2 d depending on the available computational resources and the dataset size.

bioengineering↗