Biology subjects

Park, S.-M.

Publications and source records attributed to Park, S.-M..

10 recordsLinked to original sources

bioRxiv2026

Estimating age and annual tooth growth rates in narwhals (Monodon monoceros) using computed tomography

The spiraling tooth of the narwhal (Monodon monoceros) has long intrigued scientists, but the growth rates of its "tusk" are poorly constrained due to the difficulty to age narwhal. To learn more about the population dynamics and life history of this exploited species, previous studies have used sectioned teeth to age narwhals and calibrate other ageing methods based on aspartic acid racemization, radiocarbon or trace element analysis. However, few such studies have been conducted due to the method's difficulty and the need to sacrifice the tooth. In this study, computed tomography (CT) was used on five teeth as a non-destructive means of estimating the age of narwhals, revealing different growth rates between erupted and embedded teeth during adolescence. Moreover, in combination with our reanalysis of all relevant literature, CT data clarify the annual growth rates of narwhal tusks, which peak before the age of 10 (~100 mm yr-1). This approach will be useful for investigating private and museum collections, and might also better constrain the use of trace element proxies in narwhal teeth and support sustainable harvest management.

zoology↗

bioRxiv2026

Title: Zero-shot automated insulin delivery for type 1 diabetes via dynamic physiology-aware reinforcement learning

Insulin therapy in type 1 diabetes requires constant dose adjustment based on blood glucose, meals, physiological states, and physical activity. This demanding self-management imposes a substantial burden and increases dosing-error risk, underscoring the need for automated insulin delivery (AID) systems that reduce user intervention. However, many current systems depend on fixed, individualized parameters and may not fully adapt to rapid or unobserved physiological changes. We developed the Dynamic Physiology-Aware Reinforcement learning Controller (DPARC), a zero-shot insulin optimizer that infers latent physiological dynamics from recent continuous glucose monitoring (CGM) and insulin-delivery history without prior personalization, carbohydrate announcements, or preset subject-specific parameters. DPARC uses a rolling 24-hour CGM and insulin-history window, but closed-loop operation can begin after 1 hour of observed data by initializing unobserved history with neutral normalized padding and progressively replacing it with observations. In silico, a single frozen DPARC policy adapted within 1 hour, improved time in range compared with a total daily insulin-conditioned reinforcement learning baseline, and approached the upper-bound performance of a fully personalized model under stochastic unannounced meals with randomized timing, carbohydrate amounts, absorption variability, and meal skipping. In supervised porcine studies under unannounced meals, DPARC maintained high time in range without manual configuration, supporting large-animal feasibility while prospective human evaluation is needed before clinical efficacy can be established. Learned latent representations correlated with physiological markers including insulin sensitivity and plasma insulin concentration, supporting physiological alignment and explanatory anchors. Collectively, these findings support DPARC as a preclinical proof-of-concept zero-shot AID framework for future supervised human evaluation.

bioengineering↗

bioRxiv2026

Transcriptomic Integration Reveals a Conserved Inflammatory--Proliferative Paradox in Acquired Resistance to Immune Checkpoint Blockade

Acquired resistance to immune checkpoint blockade (ICB) is increasingly recognized as an active adaptive process. However, prior studies have typically focused on individual tumor models, limiting the ability to distinguish conserved mechanisms from model-specific observations. Here, we integrated four independent transcriptomic datasets of acquired ICB resistance, spanning human non-small cell lung cancer (NSCLC) biopsies, murine CT26 colorectal tumors, organoid-derived murine NSCLC tumors, and EMT6 breast cancer cells. Differential expression analysis was performed within each dataset, followed by an intersection-based consensus approach to identify reproducible resistance-associated programs. Contrary to the conventional cold tumor paradigm, acquired-resistance tumors consistently maintained interferon-{gamma} response and innate immune signaling while simultaneously activating cell-cycle programs and constitutive KRAS signaling signatures across all four models. We term this an apparent inflammatory-proliferative paradox: the persistence of IFN-{gamma}-driven inflammatory signatures, canonically associated with productive antitumor immunity, in tumors that have escaped immune control. Notably, this program was retained in immune-depleted organoid and cell-line models, supporting a tumor-cell-associated component maintained independently of the immediate immune microenvironment. Transcription factor activity inference identified a conserved regulatory backbone linking interferon-associated regulators (STAT2, IRF2) with proliferation drivers (E2F4, TFDP1) and suppression of lineage-specifying factors (HNF4A, EGR1). Integrated network analysis resolved these signals into three reinforcing modules, namely hyper-proliferative outgrowth, active inflammatory adaptation, and lineage identity loss. This architecture provides a systems-level framework for prioritizing combination strategies that simultaneously address interconnected resistance axes.

bioinformatics↗

bioRxiv2026

Model-Driven Hybrid AI Framework for End-to-End Autonomous Decision-Making in Drug Development

Decision-making in drug development spans heterogeneous stages from molecular design to clinical optimization, yet computer-aided workflows across stages remain fragmented, limiting traceability from evidence-derived clinical questions to simulation scenarios and decision endpoints. We present a model-driven hybrid AI framework for end-to-end decision support that treats PICO (Participants, Intervention, Comparison, Outcome) as a machine-actionable specification to define scenarios and align endpoints across modules. Given a PICO-defined clinical question and a drug SMILES, the framework infers PBPK-ready parameters from structure, generates exposure via mechanistic PBPK simulation, and links exposure outputs to clinic-facing therapeutic drug monitoring (TDM) decision support using an open-source NLME-based Clinical Pharmacokinetic Consultant Service (CPCS). We additionally implement an LLM-based PICO extraction layer to structure clinical abstracts into P/I/C/O elements for systematic evidence ingestion. Results include backbone verification for DTI-guided intrinsic clearance prediction under fold-error bounds and PBPK trajectory plausibility, and external validation of the hybrid AI-PBPK pipeline showing substantial dispersion for several endpoints. An input-controlled comparison between PhysioSim and PK-Sim on N = 9 drugs yields similarly degraded accuracy under identical AI-predicted inputs, suggesting upstream parameter quality as a dominant constraint. For evidence ingestion, prompting-only PICO extraction achieves F1 of 0.687 on EBM-NLP and 0.552 on TB-PICO. For clinical utility, CPCS improves TDM performance over a PKS baseline on phenobarbital and vancomycin, reducing MAPE by approximately 26-60% depending on configuration. Overall, the framework provides a modular, traceable blueprint linking evidence-defined questions to exposure simulation and TDM decision support while preserving mechanistic interpretability. Future work will integrate outcome-aligned time-to-event modeling and federated learning to close the PICO-to-outcome loop and improve parameter inference under multi-site governance.

pharmacology and toxicology↗

bioRxiv2025

The Lipidome Landscape of Amiodarone Toxicity: An in vivo Lipid-centric Multi-Omics Study

Amiodarone is an effective therapy for arrhythmias, its prolonged management may lead to significant adverse drug reactions. Amiodarone-induced hepatoxicity is described by phospholipidosis, hepatic steatosis, cholestatic hepatitis, and cirrhosis. However, the systemic and hepatic lipidome disturbances and underlying toxicological mechanisms remain comprehensively elucidated. Untargeted lipidomics were utilized to analyze serum and liver samples from the rats orally administered a daily dose of amiodarone of either 100 or 300 mg/kg for one week. Changes in the expression of hepatic lipid-related genes were also examined utilizing transcriptomics. We found a higher magnitude of lipidome alterations in the 300 mg/kg than those in the 100 mg/kg groups. Treated animals showed elevated abundances of phosphatidylcholines, ether-linked phosphatidylcholines, sphingomyelins, and ceramides, and decreased levels of triacylglycerols, ether-linked triacylglycerols, and fatty acids. We also found 199 lipid-related differentially expressed hepatic genes between the 300 mg/kg group versus controls, implying lipid metabolism and signaling pathways disturbances. Specifically, elevation of serum phosphatidylcholines and ether-linked phosphatidylcholines, as well as hepatic bismonoacylglycerophosphates were associated with reduced expression of phospholipase genes and elevated expression of glycerophospholipid biosynthesis genes, possibly driving phospholipidosis. Perturbations of sphingolipid metabolism might also be the key events for amiodarone-induced toxicity. Alterations in gene expression levels related to lipid storage and metabolism, mitochondria functions, and energy homeostasis were also found. Collectively, our study characterized the sophisticated perturbations in the lipidome and transcriptome of amiodarone-treated rats and suggested potential mechanisms responsible for amiodarone-induced hepatotoxicity.

pharmacology and toxicology↗

bioRxiv2023

Neural computation in the brainstem for visceral sensation-driven haemodynamics

The brainstem serves as an intermediary processor of haemodynamic sensations via nucleus tractus solitaries (NTS) in regulating circulatory system. After sensing visceral inputs, the NTS relays information to efferent pathways to modulate peripheral viscera. However, the neural computation mechanism underlying how the NTS processes viscerosensory input remains unknown. Here, we show the computational principles embedded inside the NTS of rats, producing haemodynamic modulation in concert. Our findings demonstrate that the collective dynamics leveraging from neuronal population within the NTS neural circuit encode input-driven haemodynamics. The NTS exhibits the neural trajectory, the dynamical trace of neural states, which is confined to low-dimensional latent space and may represent haemodynamic perturbations. Surprisingly, by normalizing neural trajectory of rats, we found the across-rat common rules for the viscerosensory-information processing by the NTS. Furthermore, the common rules allowed to identify inter-subject variable haemodynamics by quantifying the computational mechanisms in neuro-haemodynamic axis. Our findings provide pioneering insights into understanding the neural computation involved in regulation of visceral functions by the autonomic nervous system.

neuroscience↗

bioRxiv2023

Reproducing Left Ventricular Twist by Mimicking Myocardial Fiber Orientation Using 3D Bioprinting-assisted Tissue Assembly

BackgroundLeft ventricular twist, an opposite rotation of the apex and base, is caused by myocardial fiber orientation, a unique structural feature of the myocardium, and contributes to the effective ejection fraction of the native heart. Reproducing this structural-functional relationship in an in vitro heart model remains challenging due to the lack of synchrony between layers when mimicking the fiber orientations of each layer. MethodsWe employed a hierarchical approach for creating multilayered and multiaxial fibers in a chamber-like structure, as follows: 3D bioprinting-assisted tissue assembly, fabrication of uniaxially aligned engineered heart tissue as a building block, and assembly of them to create a myocardial fiber orientation in a chamber-like structure. ResultThe EHT module confirmed uniaxial alignment and cardiac functions such as contractility and electrophysiological properties. By fabricating the assembly platform by 3D bioprinting, it is possible to guide building blocks in various directions as intended, confirming the versatility of this method. The assembly platform allows structural and functional synchrony of assembled tissues while controlling and maintaining predefined cellular alignment. Furthermore, various shapes and sizes of EHT modules and assembly platform were fabricated for mimicking myocardial fiber orientation in a chamber-like structure. The resulting structure exhibited three layers and three orientations representing myocardial fiber orientation. Moreover, the left ventricular twist was confirmed by measuring basal and apical rotations. ConclusionsRecapitulation of the microscale structure of the left ventricle enabled the identification of information not discernible from the existing macroscale structure. This understanding of the structure-function relationship of the heart can provide insights into the mechanisms underlying cardiac structure, function, and related diseases. Furthermore, the versatility of the 3D bioprinting-assisted tissue assembly allows for the creation of organs and tissue collections with complex structural and functional features by fabricating and assembling modules that meet the specific requirements of target tissues and organs.

bioengineering↗

bioRxiv2023

Comparative study of the mechanism of natural compounds with similar structures using docking and transcriptome data for improving in silico herbal medicine experimentations

Natural products have successfully treated several diseases using a multi-component, multi-target mechanism. However, a precise mechanism of action has not been identified. Systems pharmacology methods have been used to overcome these challenges. However, there is a limitation as those similar mechanisms of similar components cannot be identified. In this study, comparisons of physicochemical descriptors, large-scale molecular docking analysis, and RNA-seq analysis were performed to compare the mechanisms of action of similar compounds and to confirm the changes observed when similar compounds were mixed and used. We propose an advanced method for in silico experiments in herbal medicine research based on the results. First, physicochemical descriptors were calculated based on the chemical structures of oleanolic acid (OA), hederagenin (HG), and gallic acid (GA). Similarities were confirmed by calculating the Euclidean, cosine, and Tanimoto distances between the descriptors. Next, the mechanisms of action of OA, HG, and GA were compared and confirmed through in silico-based systems pharmacology analysis using the BATMAN-TCM platform. The proteins interacting with the three compounds were verified through large-scale molecular docking analysis using the druggable proteome. Finally, a drug response transcriptome study was performed using OA, HG, GA, and a combination of OA and HG (COH) with similar structures.A comparison of physicochemical descriptors confirmed that OA and HG were very close. In particular, the two compounds showed a concordance rate of > 99% at cosine and Tanimoto distances. The systems pharmacology analysis results confirmed that OA and HG shared more than 86% of their predicted target proteins and differed only in GA. Systems pharmacology analysis revealed that OA and HG share the mechanisms of cardiac muscle contraction, oxidative phosphorylation, and non-alcoholic fatty liver disease. In a molecular docking analysis of the 50 major druggable proteins, OA and HG shared 38 proteins, while GA shared a few with proteins derived from the other two compounds. In addition, OA and HG were confirmed to act on gonadotropin-releasing hormone (GnRH) secretion, type 2 diabetes mellitus, cholinergic synapses, and calcium signaling pathways, and docking analysis visualization confirmed that the two components interact at the same site. RNA-seq analysis also showed that the differentially expressed genes and pathways derived from OA and HG were similar, and it was confirmed that COH had similar results to OA and HG. Our study has three novel findings. First, an advanced network pharmacology research method was suggested by partially presenting a solution to the difficulty in identifying multicomponent mechanisms. Second, a new natural product analysis method was proposed using large-scale molecular docking analysis. Finally, various biological data and analysis methods were used, such as in silico system pharmacology, docking analysis, and drug response RNA-seq. The results of this study are meaningful in that they suggest an analysis strategy that can improve existing systems pharmacology research analysis methods by showing that natural product-derived compounds with the same scaffold have the same mechanism.

bioinformatics↗

bioRxiv2021

CXCR4+ Treg cells control serum IgM levels and natural IgM autoantibody production by B1 cells in the bone marrow

Regulatory T (Treg) cells represent a specialized lineage of suppressive CD4+ T cells, whose functionality is critically dependent on their ability to migrate to, and dwell in the proximity of cells they control. Here we show that continuous expression of the chemokine receptor CXCR4 in Treg cells is required for their ability to accumulate in the bone marrow (BM). Induced CXCR4 ablation in Treg cells led to their rapid depletion and consequent increase in mature B cells, foremost B-1 subset, observed exclusively in the BM without detectable changes in plasma cells or hematopoietic stem cells or any signs of systemic or local immune activation elsewhere. Dysregulation of BM B-1 B cells was associated with a highly specific increase in IgM autoantibodies and the total serum IgM levels. Thus, Treg cells control autoreactive B-1 B cells in a CXCR4-dependent manner. These findings have significant implications for understanding of regulation of B cell autoreactivity and malignancies.

immunology↗

bioRxiv2018

Small-molecule targeting of MUSASHI RNA-binding activity in acute myeloid leukemia

The MUSASHI family of RNA binding proteins (MSI1 and MSI2) contribute to a wide spectrum of cancers including acute myeloid leukemia. We found that the small molecule Ro 08-2750 (Ro) directly binds to MSI2 and competes for its RNA binding in biochemical assays. Ro treatment in mouse and human myeloid leukemia cells resulted in an increase in differentiation and apoptosis, inhibition of known MSI-targets, and a shared global gene expression signature similar to shRNA depletion of MSI2. Ro demonstrated in vivo inhibition of c-MYC and reduced disease burden in a murine AML leukemia model. Thus, we have identified a small molecule that targets MSIs oncogenic activity. Our study provides a framework for targeting RNA binding proteins in cancer.

cancer biology↗