Biology subjects

Lee, I.

Publications and source records attributed to Lee, I..

At least 19 recordsLinked to original sources

bioRxiv2026

Biohybrid toggle switches: leveraging bistability in muscle-actuated machines

Tissue-engineered skeletal muscle offers robotics unique advantages in energy density, adaptability, and self-repair, but current biohybrid machines couple muscle directly to artificial skeletons, limiting net force output and requiring continuous stimulation maintain mechanism positions. Here, we introduce a biohybrid bistable switching mechanism that converts antagonistic muscle contractions to snap-through actuation of an elastomeric four-bar flexure, maintaining stable positions without energy consumption. We decouple passive muscle pre-tensioning from flexure operation with thin-film elastomer resistance bands, enabling fine tuning of spring stiffness and force to match muscle mechanical output envelopes. Leveraging mesoscale additive manufacturing, spin coating, and replica molding techniques, components are integrated into a millimeter-scale linear stage to axially align muscles with artificial components. Our prototype achieves reversible toggling between bistable states with a 130 ms switching time, establishing skeletal muscle actuators as viable electromechanical switches, and broadening biohybrid robotics towards novel applications such as grippers, valves, and mechanical computers.

bioengineering↗

bioRxiv2026

LoGoPPI enables fast and accurate protein protein interaction mapping at scale

Graph-based protein function analysis is powerful, but protein-protein interaction (PPI) networks exist for only a small fraction of animal and plant genomes. We present LoGoPPI, which infers PPIs from sequence by combining bi-encoder global protein representation with local residue-level late interaction. LoGoPPI matches or exceeds state-of-the-art PLM-based cross-encoders while achieving orders-of-magnitude faster inference, up to ~1,500-fold, and its local branch provides residue-level signals associated with interaction interfaces and structurally flexible regions. This efficiency enables practical proteome-wide PPI reconstruction at scales prohibitive for cross-encoder models, potentially extending interactome mapping to tens of thousands of animal and plant species. LoGoPPI thus provides a scalable framework for comparative and functional analysis of protein networks across diverse taxa.

bioinformatics↗

bioRxiv2026

Human body-site-stratified fungal genomic catalogs enhancing human mycobiome analysis

Current human mycobiome profiling relies largely on marker-based methods with limited sensitivity, whereas genome-based approaches require computationally intensive alignment against broad fungal references. Here, we present MycobiomeDB, a unified fungal genome catalog, and derive body-site-specific catalogs for the gut, oral cavity, skin, and vagina through metagenome-guided stratification. Using these catalogs, we developed site-specific MycoProfilers that improved taxonomic recall and better distinguished disease-associated mycobiomes from healthy controls than existing methods.

microbiology↗

bioRxiv2026

How visual scenes recruit spatial maps: threshold-gated activation in CA3 versus continuous integration in CA1

The hippocampal cognitive map is constructed from environmental sensory inputs, but the conditions that recruit a spatial map from a quiescent baseline remain unclear. Using a parametric virtual reality paradigm with simultaneous recordings from CA1 and CA3 in rats, we demonstrate that these subregions follow fundamentally different rules. CA1 place cells exhibited robust selectivity under sparse conditions, continuously updating their tuning as landmarks accumulated. Conversely, CA3 remained quiescent, its recruitment nonlinearly gated until landmarks formed a spatiotemporally integrated ensemble, governed by ensemble density and configuration rather than absolute count. Naturalistic backgrounds facilitated early activation, eliciting stepwise recruitment as scene complexity increased. These transitions paralleled LFP dynamics shifting from sensory decoupling (reduced fast-gamma coupling) to global network engagement (increased slow-gamma coupling). Together, these findings establish a circuit-level framework wherein CA3 operates as a nonlinearly gated scene constructor and CA1 as a continuous landmark integrator to build and maintain the cognitive map.

neuroscience↗

bioRxiv2026

Coexisting but dissociable place and spatial view codes in the primate hippocampus

In rodents, hippocampal activity is classically described in terms of position-anchored place cells, whereas nonhuman primate work has emphasized spatial view cells, in which activity depends on the region of space being viewed. This contrast has fueled a longstanding debate over whether primate and rodent spatial codes are biologically distinct. We recorded single units in the hippocampus of macaques as they navigated linear tracks in two visually distinct contexts, intermittently making context-dependent object choices. Using a generalized linear model to dissociate position coding from spatial view coding, we identified three coexisting populations in the macaque hippocampus: Position cells, View cells, and Conjunctive cells encoding both variables. Position cells were prevalent and showed rodent-like properties--localized firing, directional tuning, and contextual remapping--even after accounting for visual sampling, whereas View cells stably encoded allocentric visual space. Population decoding confirmed this dissociation at the ensemble level. Robust position codes thus coexist with, yet remain dissociable from, view codes in the primate hippocampus, suggesting that the apparent scarcity of place-like responses in earlier primate work reflects methodological and task-related differences rather than a fundamental species-level divergence.

neuroscience↗

bioRxiv2026

Integrated multi-omics profiling of gut bacterial extracellular vesicles links cargo composition to host transcriptional responses

Bacterial extracellular vesicles (bEVs) enable gut microbiota to deliver bioactive cargo to host cells, yet the specific bacterial producers have not been systematically identified. Here, we profiled stool-derived bEVs from healthy individuals using metaproteomic profiling, revealing that vesiculation is widespread across gut bacterial phyla. We selected a subset of vesiculating species, showing that bEVs localize to distinct tissues in vivo, including extraintestinal sites such as lung, liver, kidney, and bone, suggesting roles beyond the gastrointestinal tract. We profiled intestinal epithelial cells and macrophages after endocytosing bEVs from various species, uncovering pronounced, cell-type-specific responses. For example, Bacteroides fragilis bEVs promote anti-inflammatory mitochondrial-telomeric regulation, while a set of commensal-derived bEVs contribute to epithelial survival and structural renewal. By combining proteomic, lipidomic, and metabolomic cargo profiling with transcriptional output, we find that specific Bacteroidota-derived protein cargo activates cytoprotective stress defenses while attenuating inflammatory signaling. Together, these findings establish a multi-layered comparative atlas of bEV composition, uptake, and host response, providing a framework for understanding bEV-mediated microbiome-host communication.

microbiology↗

bioRxiv2026

Chromatin accessibility–transcription coupling is dynamically reconfigured during Arabidopsis embryo maturation

Chromatin accessibility (CA) is widely profiled alongside transcription because it is generally associated with transcriptional activity. How directly the two layers are linked remains unclear. We addressed this in Arabidopsis thaliana embryos. We profiled CA and transcription across the torpedo, bent-cotyledon, and mature-green stages using stage-resolved nicking enzyme-assisted sequencing (NicE-seq) and nanopore RNA sequencing. To test whether their relationship is genetically regulated, we also analyzed ld-1, a mutant of LUMINIDEPENDENS (LD) linked to global transcriptional repression. Transcriptomes formed discrete stage-specific programmes, but CA did not. Accessibility at both promoters and transcription-factor motifs largely followed a common developmental trajectory that peaked at the bent-cotyledon stage and declined thereafter, rather than mirroring module-specific expression patterns. Loss of LD preserved promoter architecture but extensively remodelled motif accessibility, increasing its concordance with module-specific expression timing and strengthening CA-transcription coupling. Chromatin accessibility and transcription are subject to partly independent regulation during embryo maturation. The ability of LD loss to strengthen accessibility-transcription concordance demonstrates that their coupling is a regulated rather than fixed property of developmental chromatin.

Developmental Biology↗

bioRxiv2026

Single-nucleus multiomics unveils malignant cellular states, regulatory architectures and microenvironmental reorganization across the G-CIMP epigenomic transition in IDH-mutant glioma

IDH-mutant gliomas stratified by glioma CpG island methylator phenotype (G-CIMP) into High (GCH) and Low (GCL) exhibit markedly divergent clinical outcomes, yet cellular and regulatory determinants of this distinction remain incompletely defined. Integrating single-nucleus RNA-and ATAC-sequencing across 18 tumor specimens from 10 patients, we resolved six malignant cellular states whose differential enrichment across G-CIMP strata delineates the GCL epigenomic transition. GCL tumors were enriched for independently prognostic Mesenchymal and Mitotic Proliferative states, driven by convergent E2F, MYC, MEF2, and NFI-family regulatory networks confirmed across chromatin, histone, and transcriptomic modalities. Pseudotime trajectory inference revealed a multifurcating developmental model from an Astrocytic-like origin, with GCL tumors gaining preferential access to proliferative and mesenchymal endpoints. A reorganized immune microenvironment and candidate therapeutic axes including CDK4/6-E2F, MYC/BET, KIF11, and NOTCH, potentially combinable with vorasidenib as an epigenomic backbone provide a translationally actionable framework for intercepting GCL progression in IDH-mutant glioma.

cancer biology↗

bioRxiv2025

Sequential development of task representation from hippocampus to prefrontal cortex supports goal-directed spatial navigation

Successful goal-directed navigation requires the coordination between the hippocampus and medial prefrontal cortex. However, it is not fully confirmed that the medial prefrontal cortex learns its spatial code from the hippocampus. To test this, we examined spatial representations of the intermediate hippocampus and medial prefrontal cortex while rats learned a spatial navigation task. Rats performed a goal-directed spatial navigation task in 2D VR to find an unmarked goal zone, and we discovered robust directional tuning of single neurons in both regions. Neural manifold analysis further confirmed population-level directional tuning in both regions, with manifolds having ring-like geometry. We found that this ring-like structure evolved after learning, in a way that the hippocampal-prefrontal manifolds converged to a shared geometry. Furthermore, this evolution of ring-like structure was preceded by the hippocampus at the trial level. It was further verified that the evolution of the ring-like structure is linked to phase locking to the hippocampal theta rhythm, particularly in the prefrontal manifolds. Our findings provide compelling evidence that spatial representations of the hippocampal-prefrontal network become aligned after learning, and also highlight the information flow from the hippocampus to the medial prefrontal cortex during this geometry synchronization.

neuroscience↗

bioRxiv2025

Redefining Parkinson's Disease by Dysregulated Genetic Networks in Distinct Cell Types

Parkinsons disease (PD) is classically linked to dopaminergic neuron loss, but emerging evidence suggests broader cellular involvement. Here we show that PD risk variants converge on distinct molecular networks across specific brain cell types, enabling stratification of patients into six subgroups: dopaminergic, oligodendrocyte progenitor cells (O), excitatory (E), dopaminergic/excitatory, dopaminergic/oligodendrocyte and other. While all subgroups exhibit motor symptoms, the E-group individuals also display more severe non-motor symptoms, including dementia, hyposmia, and REM sleep behavior disorder. The O-group individuals exhibit reduced myelin integrity, as demonstrated by diffusion tensor imaging, implicating NRG6 (formerly C1orf56), a previously uncharacterized high-risk PD gene. We show that NRG6 encodes a conserved epidermal growth factor-like domain structurally and functionally analogous to neuregulin-1, which is critical for oligodendrocyte development and myelination. These findings redefine the cellular architecture of PD vulnerability and identify neuregulin-like signaling in oligodendrocytes as a potential contributor to non-motor symptoms.

neuroscience↗

bioRxiv2025

Translating clinical gene sequencing into a foundational representation of tumor subtype

While gene sequencing is routine in cancer care, translating sequences into treatment decisions remains a challenge. Here we introduce MutationProjector, an AI foundation model that transforms tumor mutation profiles into a compact representation of cancer subtype, with broad implications for diagnosis and therapy. MutationProjector is pre-trained by integrating genomic alterations from >30,000 tumors with extensive molecular knowledge, yielding a model that accurately reconstructs held-out genetic profiles (demonstrating strong generalization) and determines subtype representations from altered molecular pathways (enabling model interpretability). We evaluate MutationProjector in independent tasks related to prediction of immunotherapy response, prediction of chemotherapy response, and classification of metastasis, recording leading performance in all areas. Each task identifies key biomarkers of interest, including KMT2A and KRAS-STK11 alterations which govern immunotherapy response.

systems biology↗

bioRxiv2025

Variational Autoencoders for Metabolomics: Data Imputation, Deconfounding, and Correlation Discovery

This protocol describes a computational approach for constructing correlation-based molecular networks from untargeted metabolomics data using MetVAE, a variational autoencoder-based framework. Complementing spectral similarity networks, it captures functional relationships re-flected in cross-sample correlations. The workflow imports metabolomics features and sample metadata, adjusts for compositionality, missingness, confounding, and high-dimensionality, esti-mates sparse metabolite correlations, and exports GraphML files for network visualization. In a hepatocellular carcinoma mouse model, it links lipid classes in high-fat-diet animals, suggesting an endogenous "auto-brewery" route to lipotoxic metabolites.

bioinformatics↗

bioRxiv2025

Domain-general neural effects of associative learning and expectations on pain and hedonic taste perception

Predictive cues significantly influence perception through associative learning. However, it is unknown whether circuits are conserved across domains. We investigated how associative learning influences perceived intensity and valence of pain and hedonic taste and whether the mechanisms that support expectancy-based modulation vary as a function of aversiveness and modality. Sixty participants were randomly assigned to receive either painful heat, unpleasant liquid saline, or pleasant liquid sucrose during fMRI scanning. Following conditioning, cues that were initially associated with low or high intensity outcomes were intermittently followed by stimuli calibrated to elicit medium intensity ratings. Learned cues modulated expectations and subjective outcomes similarly across domains. Consistent with this, the orbitofrontal cortex exhibited domain-general anticipatory activation. The left anterior insula mediated domain-general cue effects on subjective intensity, whereas the thalamus mediated cue effects on subjective valence. Notably, these two regions were nearly identical to those previously implicated in mediating cue effects on pain (nearly identical to those previously implicated in cue effects on mediating pain (Atlas et al., 2010). Pain specificity was evident when we measured variations in stimulus intensity, whether we used univariate or multivariate approaches, but there was minimal evidence of specificity by modality or aversiveness when we examined cue effects on medium trials. These findings suggest that shared neural circuits mediate the effects of learned expectations on perception, linking pain with other areas of affective processing and perception across domains.

neuroscience↗

bioRxiv2025

Pre-Clinical Evaluation of a Novel Immunomodulator: a potential immunotherapy for coronaviral disease

At present, no unifying animal study to quantify the effect of PI on innate immunity in animals has been reported. This study quantifies the effect of PI in mice and THP-1 cells using a multi-disciplinary approach employing flow cytometry to measure immune function and mass spectroscopy to identify metabolomic profiles of treated cells. For proof of principle, a case study was conducted to examine the potential benefit of administering PI to improve outcomes of a feline leukemia-positive cat that also contracted the effusive form of FIP. Our results indicate that quantifiable cellular and metabolic markers result from PI treatment and can be used to establish a platform measuring the efficacy of PI in modulating the innate immune system.

cell biology↗

bioRxiv2025

A high-quality genomic catalog of the human oral microbiome broadens its phylogeny and clinical insights

Understanding the taxonomic and functional diversity of the human oral microbiome requires comprehensive genomic catalogs. We present the human reference oral microbiome (HROM), with 72,641 high-quality genomes from 3,426 species, including 2,019 novel species, significantly improving metagenomic sequence read classification over existing catalogs. Notably, HROM unveils 1,137 novel candidate phyla radiation (CPR) species, establishing Patescibacteria as the most prevalent phylum in the oral microbiota, with oral Patescibacteria forming a distinct clade from environmental Patescibacteria. We also identify an oral CPR subclade associated with periodontitis, which complement Porphyromonas gingivalis in predicting the disease. Finally, comparing HROM with reference genomes of the human gut microbiome reveals significant taxonomic and functional divergence between the oral and gut microbiomes. We identify 42 ectopic oral species and demonstrate that their relative abundance in gut microbiota is predictive of intestinal, cardiovascular, and liver diseases, highlighting the clinical importance of the oral microbiota in systemic disorders.

microbiology↗

bioRxiv2025

Minimum biomechanical energy expenditure predicts upper-limb motor strategies in individuals with limb loss

Traditional models of upper-limb motion represent observed motor behaviors as the solution to an optimization problem defined over a cost function. However, these traditional formulations are computationally expensive and it is unclear if they extend to individuals with non-standard anatomy (such as those with upper-limb loss). Goal: We propose an optimal path planning framework that leverages musculoskeletal modeling to generate motor strategies during unconstrained, upper-limb movement. Methods: We validate this framework against upper-limb trajectories measured from a 3D target acquisition task and compare performance against multiple models of upper-limb motion previously presented in literature. Results: When compared to measured upper-limb trajectories, the proposed method generates upper-limb paths with significantly less geometric error than alternative methods (p < 0.001). Significance: Our approach provides a method for upper-limb motion planning that is easily adaptable to non-standard anatomies and computationally efficient enough for prosthesis control applications. Conclusions: The proposed path planning framework provides accurate motor strategy prediction for individuals both with and without upper-limb loss.

bioengineering↗

bioRxiv2025

Lambda N as a model substrate for studying the mechanism of Escherichia coli ATP-dependent protease Lon as a regulatory enzyme.

As an ATP-dependent protease, the quality control functions of Lon have been extensively studied and reviewed in the literature. By contrast, very little research has been conducted to investigate Lons physiological functions and its mechanism as a regulatory protease. In this manuscript, we provided a survey of literature and data to convey that the lambda N ({lambda}N) protein is a suitable Escherichia coli Lon (ELon) substrate for studying the role played by Lon in regulating an RNA transcription process. For proof of principle, we demonstrated that the minimal component of the RNA transcription complex containing RNA polymerase (RNAP) and the {sigma} factor can inhibit {lambda}N degradation by ELon through SDS-PAGE, and the carboxyl-terminal of {lambda}N is important for Lon competing with RNAP interaction. Using negative stain electron microscopy, we obtained structural evidence to show that {lambda}N lacking the carboxyl-terminal flanked by residues 99-107 interacted with ELon differently than full-length {lambda}N. Taken together, the activity and EM data provide a starting point for performing a physiological enzymology study on the contribution of ELon toward RNA transcription.

biochemistry↗

bioRxiv2025

Fast and flexible minimizer digestion with digest

Minimizer digestion is an increasingly common component of bioinformatics tools, including tools for De Bruijn-Graph assembly and sequence classification. We describe a new open source tool and library to facilitate efficient digestion of genomic sequences. It can produce digests based on the related ideas of minimizers, modimizers or syncmers. Digest uses efficient data structures, scales well to many threads, and produces digests with expected spacings between digested elements. Digest is implemented in C++17 with a Python API, and is available open-source at https://github.com/VeryAmazed/digest.

bioinformatics↗