Biology subjects

Lin, S.

Publications and source records attributed to Lin, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Label-free volumetric dynamic atlas of fresh mouse organs using dynamic full-field optical coherence tomography

Dynamic full-field optical coherence tomography (D-FFOCT) enables label-free visualization of tissue microarchitecture and endogenous intracellular dynamics. Here, we establish a volumetric atlas of ten fresh mouse organs using a mobile D-FFOCT imaging platform. The system enables large-area imaging of approximately 1 cm2 within 15 min with 1,024 frames per field of view, achieving 0.75-{micro}m lateral resolution and imaging depths of up to 200 {micro}m. Large-area mosaics, three-dimensional reconstructions, and comparisons with conventional histology reveal diverse organ-specific tissue architectures. Beyond morphological characterization, quantitative analysis of dynamic intensity and mean fluctuation frequency across nine organs reveals distinct dynamic signatures and tissue heterogeneity. Longitudinal imaging of brain, liver, and intestinal tissues further demonstrates organ-dependent temporal changes in endogenous dynamics following excision. Finally, we publicly release raw D-FFOCT imaging datasets from all ten organs, providing an open resource for quantitative tissue characterization, algorithm development, and reproducible research. Together, this study provides a comprehensive, openly accessible optical atlas of fresh mouse organs, establishing a foundation for systematic tissue characterization and advancing the broader application of D-FFOCT in biomedical research.

biophysics↗

bioRxiv2026

StainAI-Web: An Interactive Platform for Whole-Slide Microglial Morphology Analysis

Microglial morphology analysis has emerged as a critical approach in neuroinflammation research. The increasing availability of high-resolution and whole-slide histology images has created strong demand for robust, scalable analytical methods capable of efficiently processing these datasets. Existing approaches may require code-based local computation or prolonged processing times for whole-slide images, while manual morphology quantification remains labor-intensive. To address these limitations, we introduce StainAI-Web, a server that performs rapid, automated microglial morphology detection across whole-slide immunohistochemistry images in the cloud. The system provides interactive grayscale preprocessing and supports ROI-based regional quantification, allowing users to define specific tissue regions and systematically compare microglial populations across regions. An interactive visualization module dynamically updates quantitative summaries of morphological distributions as ROIs are selected or modified. Analysis results, including structured JSON files and morphology-annotated overlays, are downloadable for downstream analysis or integration into external computational pipelines. These integrated features provide an accessible framework for efficient and reproducible whole-slide microglial morphology analysis. The StainAI-Web is available at https://stainaiviewer.azurewebsites.net/.

bioinformatics↗

bioRxiv2026

NeoOmni Atlas: an integrated database for pan-cancer neoantigen derived from proteogenomics analysis

Tumor neoantigens derived from diverse molecular aberrations are important targets for personalized cancer immunotherapy, but existing resources incompletely integrate mass-spectrometry-supported canonical and non-canonical antigen evidence across cancer types. NeoOmni Atlas is an interactive pan-cancer T-cell epitope resource, built via proteogenomic reanalysis of 7,561 immunopeptidomics datasets comprising 511 million raw spectra from 2,341 samples across 26 cancer types. The analysis yielded 872,414 eluted canonical peptides and 198,961 non-canonical peptides spanning 18 molecular alteration classes, along with 759,386 unique ligands presented by 322 human leukocyte antigen (HLA) alleles. Relative to existing repositories, NeoOmni Atlas covers a broader range of cancer types, includes 60 additional HLA alleles, and catalogs 369,781 novel ligands. A computational tumor-specific filtering framework identified 316,126 tumor-associated epitopes and 61,231 tumor-specific epitopes. Tumor-specific epitopes showed stronger predicted binding affinity than tumor-associated epitopes for HLA class I A and B alleles, and 65 conserved tumor-specific epitopes recurrent across more than 10 tumor types represent potential shared antigen candidates. NeoOmni Atlas also prioritizes candidate tumor-associated antigens including PMEL (gp100:209-217) and CSPG4, and supports browsing, searching, and downloading by cancer type, human leukocyte antigen allele, peptide sequence, and source event. NeoOmni Atlas is freely accessible at http://modinfor.com/NeoOmni-Atlas/.

bioinformatics↗

bioRxiv2026

Deciphering the gene-centric spatial dynamics across tissue and time with GeneSPOT

Spatial omics technologies provide unprecedented opportunities to investigate how molecular programs are organized within tissues. However, existing analytical frameworks primarily focus on identifying cellular states and characterizing differentially expressed molecular features across these states. The relational spatial organization of molecular features and its dynamics across biological contexts therefore remain largely unexplored. Here, we present GeneSPOT, a gene-centric computational framework for characterizing the spatial relationships among genes and quantifying their reorganization across biological conditions. Although gene-centric in design, GeneSPOT generalizes to other molecular features. Across diverse applications, GeneSPOT resolved tissue structures, integrated paired and unpaired spatial multi-omics data, unveiled the spatial convergence of gene programs during embryonic development, and identified aging-associated spatial reorganization among genes whose expression did not change significantly with age. GeneSPOT thus complements conventional cell-centric analysis, providing a framework for deciphering how spatial relationships among molecular features are reorganized across biological conditions and over time.

bioinformatics↗

bioRxiv2026

Deciphering Mechanistic Signatures in Drug-Drug Interactions with Dual Topology Graphs

Drug-drug interactions (DDIs) represent a critical challenge in drug development and clinical practice, as they can lead to severe adverse effects, including toxicity and reduced therapeutic efficacy. Deep learning methods have shown promise in large-scale, rapid DDI prediction; however, current approaches suffer from significant limitations in providing mechanistic insights into these interactions. Here, we propose DualTopoDDI, a dual-topology-enhanced interpretable deep learning model for DDI prediction. We applied DualTopoDDI to predict 9.2 billion potential interactions among approved drugs, achieving 97.99% high-confidence predictions. The model demonstrates molecular-level interpretability, identifying key substructures responsible for drug actions in terms of both atom-centric and bond-centric views. DualTopoDDI also excels in elucidating particular DDI toxicity mechanisms, which is validated by its successful explanation of the controversial cardiac toxicity in two COVID-19 drug combination regimens at the time. Evaluations across 11 benchmark datasets demonstrates that DualTopoDDI not only achieves state-of-the-art performance but also showcases robust generalizability and good interpretability. Overall, DualTopoDDI offers a powerful, interpretable tool for understanding and predicting drug-drug interactions, providing critical insights for drug safety and design.

bioinformatics↗

bioRxiv2026

ssJSD: A fusion of sparsity and spatial information for HiC single-cell clustering

Single-cell high-throughput chromatin conformation capture (scHiC) enables profiling three dimensional genome architecture at cellular resolution, providing insights into cell-to-cell variability and cellular functions. Recent frameworks utilize spatial interaction patterns to derive dissimilarity measures for downstream tasks such as cell clustering. However, the inherent sparsity and ultra-high dimensionality of scHiC contact matrices pose significant challenges. A central hurdle is that existing measures typically treat all zeros without distinction, failing to differentiate biologically meaningful structural zeros (SZs) from technical dropouts. Here, we introduce ssJSD (spatial and sparsity informed Jensen-Shannon Divergence), a computational framework designed to explicitly account for scHiC-specific sparsity patterns. By integrating band-wise contact frequency profiles with SZ-induced sparsity matrices, ssJSD leverages both spatial interaction patterns and biological absence of contacts. We adopted two complementary integration strategies: an early fusion approach that concatenates information into a single representation, and a late fusion approach that integrates JSD-based dissimilarities through diverse averaging methods. Through simulations and applications to human cell lines and prefrontal cortex data, we demonstrate that ssJSD improves clustering accuracy and effectively distinguishes cell types. Our study indicates that integrating SZ patterns is important for accurately quantifying cell-to-cell variability in 3D genomics.

genomics↗

bioRxiv2026

A Full Spectrum Fibroblast Profiling Panel Reveals Phenotypic and Metabolic Heterogeneity of Cancer-Associated Fibroblasts at the Single-Cell Level

Cancer-associated fibroblasts (CAFs) exhibit extensive transcriptional and phenotypic heterogeneity and are highly plastic. Routine approaches to resolving CAF heterogeneity at the single-cell and protein levels are lacking, making it difficult to assess changes in CAF state in response to perturbations. Here, we present an optimized 28-marker spectral flow cytometry panel that uses unsupervised clustering to resolve CAF states across in vivo and in vitro models. Using this new fibroblast profiling panel (FPP), we show that LRRC15+ CAFs in murine pancreatic ductal adenocarcinoma are phenotypically distinct from canonical SMA+ myofibroblasts, rather than representing a subset of this population. We find that marker expression is graded rather than discrete across clusters, indicating that CAF identity spans a continuum rather than fixed states. Metabolic profiling further shows that hypoxic CAFs were enriched for glucose uptake and an inflammatory phenotype. Using in vitro models, we further demonstrate how the FPP can resolve plasticity outcomes. In a 3D organoid co-culture model, direct contact between tumor cells and fibroblasts, but not organoid-derived factors alone, induces a stem-like CAF state also observed in vivo. In 2D, TGF{beta} and inflammatory/hypoxic stimuli each enrich distinct, pre-existing CAF states rather than generating new ones. These findings establish our FPP as a practical method for resolving fibroblast heterogeneity at the protein level in cancer and beyond.

cell biology↗

bioRxiv2026

Discovery of an NRAS isoform and activation-state selective macrocyclic peptide

Macrocyclic peptides have gained increased attention amid claims they are a "Goldilocks" therapeutic modality that can encode the selectivity of a biologic in a footprint close to that of a small molecule. Here we attempt to find a peptide that binds selectively to NRAS, sparing HRAS and KRAS, while accessing the cytosol via passive cell permeability. To do so, we combine subtractive affinity selection with mRNA display to identify Compound 1, an 11mer macrocyclic peptide which binds NRAS at a novel allosteric site between Helix 3 and Helix 4 of the GTPase domain. Compound 1 has total isoform selectivity and can be tuned to achieve activation-state selectivity with a single amino acid change. While it has preferential affinity for oncogenic NRAS-specific mutations, it does not inhibit NRAS function or achieve passive membrane permeability.

cancer biology↗

bioRxiv2026

Microbial Metabolites Potentiate MAIT Cell Anti-Tumor Immunity Against Solid Tumors

Mucosal-associated invariant T (MAIT) cells sense riboflavin metabolites through the monomorphic antigen-presenting molecule MR1, providing a unique opportunity to therapeutically mobilize a broadly shared T cell population without genetic engineering. Here, we show that the highly potent microbial metabolites, 5-OP-RU and 5-OE-RU, can be exploited as pharmacologic precision immune activators to drive human MAIT cell responses against solid tumors. Ligand stimulation in human co-culture systems elicited robust MAIT cell cytotoxicity, inflammatory cytokine secretion, and transcriptional states transformation. In vivo administration of riboflavin ligands 5-OP-RU significantly suppressed tumor growth in xenograft liver cancer models. Metabolite-driven MAIT activation also reprogrammed the local immune landscape, enhancing effector function and overcoming features of the immunosuppressive niche by markedly eliminating tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) within the tumor microenvironment (TME). These findings reveal that microbial riboflavin metabolites can power and redirect MAIT cells to solid tumors, establishing MR1-metabolite signaling as a tractable therapeutic axis for liver cancer and other solid malignancies.

cancer biology↗

bioRxiv2026

Heparan Sulfate Controls Nanoscale Assembly of GPC3-Wnt Receptor Complexes

Glypican-3 (GPC3) is a heparan sulfate proteoglycan that is highly expressed in hepatocellular carcinoma and promotes tumor progression through Wnt3a/{beta}-catenin signaling. However, how the nanoscale organization of GPC3 at the cell surface controls signaling remains unclear. Here, we combined nano-resolution MINFLUX imaging, single-molecule tracking, and functional assays to define the spatial architecture and dynamics of GPC3 on hepatoma cells. We found that GPC3 exists as both single molecules and nanoscale clusters and switches between confined and free diffusions on the plasma membrane. Heparan sulfate (HS) chains create nanoscale corrals that limit GPC3 movement, whereas removal of HS increases diffusive heterogeneity and disrupts confinement. Wnt3a stimulation induces the formation of higher-order GPC3 assemblies and enhances {beta}-catenin signaling, while loss of HS markedly reduces this response. MINFLUX DNA-PAINT further revealed that HS chains orchestrate the spatial distribution of Wnt3a and promote its association with the Wnt receptor, Frizzled-1, an essential step for pathway activation. Collectively, these findings reveal that HS controls the nanoscale organization and dynamics of GPC3 to promote Wnt receptor assembly and efficient {beta}-catenin signaling in hepatoma cells.

cell biology↗

bioRxiv2026

MAERM: Predicting Enzyme-Reaction Matching Relationships with a Mixed-Attention Model

Harnessing enzyme specificity requires a thorough understanding of enzyme promiscuity, which determines enzymes catalytic scope; however, measuring this scope still relies heavily on labor-intensive analytical approaches. While data-driven approaches have emerged to predict the catalytic scope of enzymes, these methods continue to face challenges such as restricted datasets and insufficient integration of enzyme structural information and reaction transformations. Here, we introduce MAERM, an innovative mixed-attention model designed to predict enzyme-reaction matching relationships. Built on our MAERM-DB, a dataset with broad coverage of validated and chemoenzymatic catalysis data, MAERM utilizes a local-global attention module to integrate multimodal enzyme information with fine-grained reaction representations, thereby predicting enzyme-reaction matching probabilities. Results show that MAERM consistently outperforms all baselines, with an average F1-score of 0.984. Notably, on challenging test samples with less than 40% sequence identity to the training set, MAERM outperforms the second-ranked model by 5.9% in F1-score. In addition, MAERM achieves the highest top-10 success rate of 51.7% on Enzyme-405 and the highest balanced accuracy of 0.697 on BioCat-547, further supporting its generalizability in enzyme screening and chemoenzymatic catalysis. Finally, MAERM can serve as an efficient scoring module. When integrated with ProteinMPNN, MAERM has successfully guided novel enzyme design for two carbonyl reduction reactions, resulting in enhanced catalytic potential for the native substrate and demonstrating broad compatibility. Overall, MAERM has the potential to reduce the experimental cost of measuring enzymes catalytic scope, facilitate enzyme design, and ultimately accelerate the design-build-test-learn cycle in enzyme engineering.

bioinformatics↗

bioRxiv2026

MOTS-c Coordinates Inter-Organellar Proteostasis for Adaptation to Chronic Metabolic Stress

Mitochondrial communication coordinates adaptive responses across organelles to sustain cellular homeostasis, a network that declines with age and contributes to loss of proteostasis. Here, we identify MOTS-c, an exercise-induced mitochondrial-derived peptide (MDP) encoded within the 12S rRNA locus, as an inter-organellar arm of the mitochondrial stress response (MSR) that links mitochondrial signaling to endoplasmic reticulum (ER) proteostasis and enables adaptation to chronic stress. Using progressive stress media (PSM), a model of gradual and multifactorial metabolic stress, we show that MOTS-c enables adaptation through a biphasic program: acutely, a reversible, ATF4-independent suppression of protein synthesis; and chronically, an ATF6-biased ER unfolded protein response (UPRER) with tempered ATF4 engagement and coordinated metabolic remodeling. Whereas mitochondrial unfolded protein response (UPRmt) pathways have been extensively characterized in acute, genetic, and sustained models of mitochondrial perturbation, this work reveals how mitochondrial communication actively engages ER proteostasis during progressive and persistent metabolic stress. By expanding proteostatic capacity while tempering terminal stress signaling, MOTS-c enables cells to withstand chronic stress. Together, these findings define a MOTS-c-dependent arm of the MSR that integrates mitochondrial communication with ER proteostasis to promote chronic metabolic stress adaptation.

Cell Biology↗

bioRxiv2026

Hypermutability of ultraconserved histone genes and its contribution to human disease

Histone genes are among the most evolutionarily conserved sequences in eukaryotes, reflecting their essential role in nucleosome architecture. Intriguingly, our analysis reveals unexpectedly high germline mutation rates in replication-dependent histone genes across vertebrates. We provide multiple lines of evidence suggesting that this hypermutability is largely driven by off-target mutagenic effects of AID/APOBEC enzymes, representing an unintended fitness cost associated with vertebrate immune system innovation. To characterize the impact of these mutations, we develop HistMTR, a paralog-aware missense constraint framework that outperforms existing tools in predicting the pathogenicity of histone mutations. Applying this framework to large-scale human genomic data, we demonstrate that both coding and regulatory histone mutations contribute to a spectrum of diseases, including developmental disorders and reproductive failure. Analyses of allele transmission ratios and the distribution of fitness effects indicate intense ongoing negative selection against deleterious histone mutations in humans. Together, these findings uncover a hidden vulnerability of ultraconserved histone genes and establish HistMTR as an important framework for characterizing the underappreciated role of histone mutations in human disease.

evolutionary biology↗

bioRxiv2026

Reinforcement learning-driven unified generative framework for multi-objective RNA codon design

Current RNA codon design methods are limited by inefficient long-sequence processing and poor generalizability, often relying on a decoupled generate-or-optimize paradigm. We introduce RNARL, a reinforcement learning-driven framework that unifies sequence generation with multi-objective optimization. RNARL directly learns to generate high-performance sequences, effectively optimizing sequences over 3,900 nucleotides and demonstrating superior performance and universality across six species and five RNA types. RNARL thus establishes an effective and generalizable framework for RNA codon design. Finally, a user-friendly web platform is freely available to facilitate its application for RNA therapeutic design.

bioinformatics↗

bioRxiv2026

Phenotype-environment matching in ground-nesting birds across and within large-scale biomes

Many animals have camouflage appearances that correspond to the habitat where they live, termed phenotype-environment matching. However, this has generally been tested in only a limited number of systems and natural settings, and often not to the relevant vision of key receivers (predators) across different spatial scales. Here, we measured the plumage attributes and putative camouflage used by ground-nesting bird species specialist to particular biomes, specifically tropical rainforest, taiga forest, dry forest, grassland, desert, and tundra from museum specimens. Digital photography and image analysis were used to quantify colour patterns to models of predator vision to understand how different colour patterns may correspond with biome type. With this information, we next created bird models that were photographed in situ in the Valdivian temperate rainforest and Patagonian grassland biomes of Chile to quantify the extent to which plumage coloration and pattern traits provide effective camouflage at different spatial scales. In general, we find that specialist ground-nesting birds express a phenotype that better matches the substrate composition and vegetation structure across large spatial scales of their own biome. This study reveals how animal camouflage works across biomes, relative to the visual system of raptor predators, and at the appropriate distance at which detection may occur.

ecology↗

bioRxiv2026

Natural genetic variation spans all predicted Rh5/Rh6 expression phenotypes in Drosophila R8 photoreceptors

Standing genetic variation is immediately available to selection, but how broad a range of phenotypes it can produce remains unknown for many traits. Here we address this question using the mutually exclusive Rhodopsin 5 (Rh5) and Rhodopsin 6 (Rh6) expression pattern in Drosophila R8 photoreceptors, a neuronal differentiation readout that permits enumeration of a finite set of predicted qualitative phenotypic changes. Wild flies and wild-derived inbred lines from the Drosophila Genome Reference Panel 2 (DGRP2) showed extensive variation in Rh5/Rh6 expression, including shifts in the relative abundance of Rh5- and Rh6-expressing R8s, Rh5/Rh6 co-expression, and loss of Rh5 or Rh6. Among 205 DGRP2 lines, we observed examples spanning all eight predicted qualitative phenotype categories. We also identified causal coding and regulatory variants, including alleles of sevenless, Rh5 and Rh6, and an intronic deletion in melted. These results show that standing natural variation can span the full predicted range of qualitative phenotypes in this system.

genetics↗

bioRxiv2026

Integrating Histology with Spatial Molecular Programs Using a Multimodal Foundation Model

Histopathological assessment remains central to cancer diagnosis and stratification, yet its mechanistic interpretation remains limited without molecular context. To address this, we developed SQUALL, a multimodal foundation model integrating histology with spatial molecular programs. For pretraining, we assembled histMol, a large-scale corpus of 1.76 billion paired histology-spatial transcriptomics spots/bins across 33 tissues and 12 platforms from 3,446 tissue sections. Following pretraining, SQUALL enables transcriptome-wide virtual biomarker profiling, prognostically relevant spatial niches discovery, and integrative disease progression modeling. Leveraging its multimodal embeddings, SQUALL identifies niches associated with tertiary lymphoid structure (TLS) maturation and ovarian cancer relapse, reconstructs molecular trajectories of breast cancer invasion across 325,112 spots, and uncovers underlying transcriptional programs. Applied to whole-slide images from 898 patients, SQUALL outperforms existing pathology foundation models in outcome prediction while enabling interpretable risk stratification. Together, these results establish spatially aligned multimodal pretraining as a new paradigm for extending molecular insights into pathology images.

bioinformatics↗

bioRxiv2026

Mirror-image mRNA display uncovers isoform-selective D-peptide macrocycles targeting a cryptic KRAS pocket

Activating KRAS mutations drive millions of cancers diagnosed worldwide,1 yet for decades this oncoprotein was deemed "undruggable", reflecting the challenge of discovering molecules capable of perturbing its complex biological functions, and of translating these discoveries into effective cancer therapeutics.2 Recent advances propelled by innovative screening have identified diverse modalities that bind at or near the switch-II pocket (SII-P) of RAS proteins, including molecular glues,3 macrocyclic peptides,4 fragment-derived small molecules,5 and approved therapies that covalently target KRASG12C.6,7 Unfortunately, resistance to approved therapies has emerged,8,9 highlighting the need for molecules that engage new or underexploited binding sites on RAS oncoproteins with mechanisms complementary to established SII-P inhibitors.10,11 Here we show that mirror-image mRNA display12 enabled the discovery of all-D macrocyclic peptide ligands targeting a cryptic RAS back pocket (CRB-P).13 These ligands engage KRAS(OFF) and KRAS(ON) with equal affinity, exploit a single-residue difference among isoforms to bind KRAS selectively, and successfully inhibit oncogenic signaling in KRAS-mutant cells through a mechanism distinct from SII-P binders. Mirror-image screening directly afforded nanomolar peptide ligands stable toward cellular proteolysis and delivered probes targeting distinct epitopes not accessible by homochiral peptide-display methods. Together, these findings establish the CRB-P as a specifically druggable and mechanistically differentiated site on KRAS with potential for combination with emerging RAS-targeting therapies and substantiate mirror-image mRNA display as a strategy for discovering stable all-D macrocyclic peptides targeting previously inaccessible epitopes on challenging targets.

cancer biology↗