Biology subjects

Yang, Y.

Publications and source records attributed to Yang, Y..

At least 271 records · Page 15Linked to original sources

bioRxiv2025

Diverse infections transcriptionally reprogram the intestinal epithelium and epithelial-immune cell interactions

The distal small intestine plays vital roles in host physiology by regulating nutrient and fluid homeostasis. Despite being impacted in Crohns disease and a major target for a range of infections, we know relatively little about the complexity of cellular responses and cell-cell communication in the ileum during infection. Single cell and spatial transcriptomics have emerged as powerful technologies to study tissue heterogeneity in the gut, but these tools have focused on the large intestine, in part due to the accessibility of this tissue for biopsies and its importance in cancer. Here we present GutPath, an atlas of over 500,000 single cells with RNA and protein expression profiles for 91 cell states in the ileum across diverse infectious archetypes. We show that GutPath accurately captures established immune responses to infection while revealing pathogen-specific responses in enterocytes. To highlight the discovery potential of this atlas, we identify a novel enterocyte cell state present during Yersinia pseudotuberculosis infection that is spatially linked to bacterial load and tissue pathology. GutPath establishes a much-needed resource for the immunology community that will accelerate the study of the transcriptional diversity of cellular landscapes in the small intestine.

immunology↗

bioRxiv2025

Physiological Arousal as a Predominant Source of Individual Differences in Functional Brain Networks

Individual differences in brain network function and organization are promising targets for fMRI-based biomarkers in precision psychiatry, yet the sources of individual variability remain poorly understood. We show that arousal, assessed by systemic low-frequency oscillation (sLFO) amplitude in the fMRI signal, accounts for a substantial portion of variance in brain network properties (e.g.: R= 0.70 for default mode network (DMN)-dorsal attention connectivity; R= -0.63 for DMN dynamics). These relationships replicated across sessions, across independent samples, and when assessed with traditional arousal indices. Critically, associations persisted after sLFO denoising, indicating a genuine brain-physiology relationship rather than hemodynamic artifact. Pharmacological manipulation showed that drug-induced sLFO-assessed arousal changes were accompanied by corresponding shifts in network connectivity and dynamics. This work identifies arousal as a prominent determinant of variability in functional brain networks, providing a new perspective on brain-based individual differences while offering an approach to measure arousal directly from fMRI data.

neuroscience↗

bioRxiv2025

Eosinophil-derived COX-2 protects against experimental colitis through the PGE2-IL-22 axis

Inflammatory bowel disease (IBD) is driven by a breakdown in immune regulation and epithelial barrier function, yet the contribution of eosinophils to this process has remained poorly defined and controversial. While eosinophils infiltrate the intestinal mucosa during both flares and remission, their role in shaping disease outcomes is unclear. Our RNA-seq analyses of colonic eosinophils isolated from dextran sulfate sodium (DSS)-treated mice revealed a significant upregulation of cyclooxygenase (Cox)-2 (gene name, Ptgs2). Eosinophil-specific deletion of Cox-2 (Ptgs2fl/fleoCre+/-) reduced IL-22 production and exacerbated DSS- and trinitrobenzene sulfonic acid (TNBS)-induced colitis, characterized by greater weight loss, higher disease activity, colon shortening, and epithelial injury. Administration of recombinant IL-22 reversed these phenotypes. Mechanistically, eosinophil-derived COX-2 enhanced IL-22 production by type 3 Innate lymphoid cells (ILC3s) through prostaglandin E2 (PGE2) signaling. Consistently, Ptgs2fl/fleoCre+/-mice exhibited reduced colonic PGE2 levels, while PGE2 analog treatment restored IL-22 production and mucosal protection. Our findings identify eosinophil-derived COX-2 and PGE2 as a critical regulator of IL-22 production during colitis, uncovering a previously unrecognized eosinophil-ILC3 crosstalk that safeguards the intestinal barrier and represents a promising therapeutic target in IBD. Significance StatementThe role of eosinophils in inflammatory bowel disease (IBD) has remained unresolved. We identify eosinophil-derived COX-2 as the main source of prostaglandin E2 (PGE2) in colitis, which drives IL-22 production by ILC3s to maintain epithelial protection. This discovery defines a protective eosinophil/PGE2/IL-22 circuit and reveals a therapeutic target to strengthen barrier defense in IBD.

immunology↗

bioRxiv2025

AI Identifies Broadly Neutralizing Antibodies from an IGHV1-69 Public Antibody Class Exerting Continued Selection Over SARS-CoV-2

R1-32-like public antibodies, characterized by shared IGHV1-69/IGLV1-40 usage, are elicited in more than 50% of individuals with COVID-19 and have been implicated in driving recurrent mutations at L452SARS2 and F490SARS2 within their convergent epitope in the SARS-CoV-2 spike receptor-binding domain. These mutations effectively mediate escape from non-affinity-matured R1-32-like antibodies with germline-like sequences. Here, we characterize four affinity-matured R1-32-like antibodies, C092, C807, BD56-104, and BD56-597, that tolerate L452SARS2 and F490SARS2 mutations. We show that this tolerance arises from residues introduced by somatic hypermutation at convergent positions across multiple CDR loops and surrounding regions, thereby creating additional contacts that reinforce epitope binding. An unusual N354SARS2 glycosylation site, which emerged in BA.2.86 and became fixed in its descendants, is linked to escape from affinity-matured R1-32-like antibodies, implying ongoing selection by this public antibody class. Using an AI model trained on integrated structural, neutralization, and binding data, we further identified ZL525, an ultrapotent R1-32-like antibody with pan-SARS-CoV-2 variant activity, including against the highly evasive KP.3 variant carrying the N354SARS2 glycosylation, and broad sarbecovirus cross-reactivity extending to SARS-CoV-1. Together, these findings show how affinity maturation enables public antibodies to adapt to viral antigenic drift, reveal their role in shaping SARS-CoV-2 antigenic evolution, and demonstrate the potential of AI-empowered strategies for discovering broadly neutralizing antibodies.

immunology↗

bioRxiv2025

A standardized and reproducible behavioral protocol to elicit visual spatial attention in mice

Understanding how neural activity gives rise to cognitive processes such as selective attention is a fundamental goal of neuroscience. An important but often overlooked advance towards this goal requires the development and sharing of rigorous and reproducible behavioral tasks across labs; this is particularly important given the recent surge in studies of cognition and perception in mice. Here, we developed a standardized training protocol for head-fixed mice to become experts in a psychometric visual contrast detection task in just 17 days. Experts detected stimuli at two distinct spatial locations for several hundred trials per day. As consecutive trials elapsed at either location, the speed, accuracy, and contrast sensitivity of visual perception improved - all hallmarks of spatial attention improving performance, as seen in primates. These findings validate the efficacy of this protocol to reveal multiple aspects of selective spatial attention in mice, establishing a rigorous and reproducible tool for the community.

neuroscience↗

bioRxiv2025

OpenStride: an inexpensive, open-source force plate actometry system for quantification of rodent motor activity and behaviour

IntroductionForce plate actometry (FPA) enables quantitative assessment of rodent motor and behavioural activity. Through tracking centre of mass at high temporal and spatial resolution, FPA can directly quantifying motor performance and various behaviours in the context of operator-independent, naturalistic movement. Previously designed systems have been expensive and non-customisable, and they are no longer commercially available. MethodsWe designed OpenStride, an open-source FPA system consisting of both hardware and software. Our system was designed to be able to be constructed with standard tools worldwide at a cost of [~]$550 USD, with modifiable hardware files, modular software, and support for both Mac and Windows operating systems. Required tools include 3D printing and acrylic laser cutting. ResultsOpenStride reliably tracks position within a 30 x 30 cm2 environment. Based on these positional data, OpenStride currently quantifies tremor, ataxia, distance, velocity, low-mobility bouts, and centre-vs-margin time, with potential to expand to additional analyses. ConclusionsOpenStride is intended to provide a valuable tool for high-throughput, inexpensive study of motor and behavioural function and dysfunction. Software and hardware files will be freely disseminated online via GitHub at the time of final publication to enable others to construct and utilise the OpenStride system.

neuroscience↗

bioRxiv2025

Multi-Timescale Neural Adaptation Failure as a Mechanistic Signature of Major Depression

Major depression (MD) is usually characterized in terms of momentary abnormalities, such as negative bias or emotional blunting, derived from paradigms that measure only the first response to a stimulus. Yet depressive experience in daily life is repetitive and cumulative. Here we test the hypothesis that a failure of neural adaptation across repetitions is a core feature of MD. Using a repeated negative-stimulus paradigm, we quantified adaptation at micro (within-trial) and macro (across-session) scales via electroencephalography in 28 MD patients and 34 healthy controls, measuring both conventional event-related potentials and large-scale traveling-wave propagation. MD patients exhibited pervasive maladaptation: neural rigidity at the micro scale (absent repetition suppression) and progressive regulatory depletion at the macro scale (escalating emotional reactivity with declining cognitive control). Across analyses, adaptation-based metrics revealed larger group differences than single-shot indices, and classifiers built on adaptation features--especially traveling-wave dynamics--outperformed single-shot models (task AUC = 0.83 vs 0.78) and generalized to two independent datasets (up to AUC = 0.91). These findings challenge conventional symptom-based characterizations of MD, establishing adaptation failure--not initial reactivity--as the mechanistically primary pathological signature. This reframing provides a unified mechanistic explanation for rumination and cognitive inflexibility while offering a standardized, low-cost biomarker requiring minimal electrodes ([≤]8 channels), enabling large-scale screening in resource-limited settings.

neuroscience↗

bioRxiv2025

Cellular interactions in the sentinel lymph node predict melanoma recurrence

Melanoma outcomes have dramatically improved over the past decade, but some patients still experience disease recurrence, particularly those who present at later stage of disease. Prior studies identified an altered immune microenvironment in the sentinel lymph node (SLN) including the presence of dysfunctional CD8 T cells and CD4 T regulatory cells (Tregs), but the spatial organization of the SLN and the interactions of individual cell types have not been extensively studied. To understand how spatial organization of immune cells in the SLN is related to outcomes, we performed spatial proteomic profiling of Stage I and II melanoma SLN to generate an atlas of cell types. Following multiplexed immunofluorescence imaging, deep learning-based segmentation and clustering, we identified 33 subsets of T cells, B cells and Tumor cells. Lymphoid region analyses established a foundational spatial map of the melanoma SLN, revealing higher regional frequencies of activated and memory CD4 T cells in Stage I SLN and consistent positioning of T cells at functional spatial locations. To better understand cellular interactions, we evaluated the immediate neighbors around each index cell using Effect Size Interaction mapping (ESI-map), a novel computational toolkit for understanding spatial interactions. Nearest-neighbor analyses across biological replicates revealed rewiring in cell-cell interacting pairs across disease progression and patient outcomes, particularly with respect to exhausted TOX+ CD8 T cells. Stage II patients who experienced disease recurrence had notable enrichment of cellular interactions between Tregs and TOX+ CD8 T cells that was associated with reduced expression of granzyme B and interferon-{gamma} that was most pronounced in exhausted CD8 T cells. Together, these data demonstrate that the specific spatial interactions between Tregs and exhausted CD8 T cells in the SLN provide a novel immune signature for understanding melanoma outcomes. One Line SummaryTreg-CD8 T cell interactions in the sentinel lymph node predict recurrence outcomes in Stage II SLN+ melanoma patients.

immunology↗

bioRxiv2025

DfE-DB: A systematic database of 3.8 million human decisions across multiple experience-based tasks

Learning from experience is central to human decision making, but progress is limited by heterogeneous task designs and fragmented data formats that hinder systematic comparison across studies. We introduce the Decisions From Experience Database (DfE-DB), a structured, open-access resource comprising 3.8 million trial-level decisions from 11,921 participants across 168 studies and 13 paradigms. By harmonizing raw behavioral data and classifying studies along 13 key design features, the DfE-DB enables quantitative comparisons previously obscured by heterogeneous task and data structures. Using the database, we show that choice tendencies--toward higher risk, expected value, or experienced mean--vary substantially across paradigms and are strongly shaped by core design features such as feed-back type, outcome structure, stationarity, and sampling. These features explain substantial cross-study variability and reveal underexplored paradigm variants. By providing a unified and extensible data infrastructure, the DfE-DB facilitates reproducible research on testing the generality of behavioral phenomena and computational models, fostering a more integrated science of decisions from experience.

neuroscience↗

bioRxiv2025

A Combinatorial mRNA Therapy for Treating Rheumatoid Arthritis and Osteoarthritis by Inhibiting Inflammation and Promoting Cartilage Regeneration

Rheumatoid arthritis (RA) and osteoarthritis (OA) are debilitating joint disorders with distinct etiologies but share common pathological features of chronic inflammation and progressive cartilage damage. Current therapeutic options are largely palliative and yet fail to achieve desired effect. In this study, we developed a combinatorial mRNA therapy using a rationally engineered lipid nanoparticle (LNP) containing a novel ionizable lipid for enhanced cartilage penetration upon intra-articular administration. This LNP co-delivers mRNAs encoding two complementary therapeutic proteins: interleukin-1 receptor antagonist (IL-1Ra) for inflammation attenuation and a C-terminal truncated derivative of angiopoietin-like 3 (ANL3) for cartilage regeneration. In pre-clinical models including collagen-induced arthritis mice, TNF-transgenic mice with spontaneous RA and destabilization of medial meniscus (DMM) surgical OA model, this "one-shot" combinatorial therapy significantly ameliorated disease severity, reduced synovitis and bone erosion, and potently promoted the regeneration of hyaline-like cartilage. Mechanistically, transcriptomic profiling of joint specimens revealed that the combinatorial therapy concurrently suppressed inflammatory pathways and extracellular matrix degradation pathways, meanwhile upregulating anabolic genes facilitating chondrogenesis. Collectively, our study establishes a versatile and synergistic mRNA therapy that offers a potent disease-modifying immunomodulatory strategy capable of addressing the long-standing unmet needs in arthritis treatment.

immunology↗

bioRxiv2025

MisTIC: Missegmented Transcript Inference Correction for Improved Spatial Transcriptomics Analysis

Imaging-based spatially resolved transcriptomics (SRT) technologies, such as 10X Xenium, MERSCOPE, and CosMx, have revolutionized our ability to study gene expression within the spatial context of tissues at single-cell resolution. The acquisition of such data relies heavily on cell segmentation algorithms, which often produce imperfect boundaries, leading to transcript misassignment. These misassignments can significantly affect downstream analyses, including cell type identification, differential expression analysis, cell-cell communication, and RNA localization. We present MisTIC (Missegmented Transcript Inference Correction), a variational Bayesian model designed to correct transcript misassignment errors without requiring resegmentation. In benchmarking analyses using synthetic data with simulated transcript misassignment, MisTIC demonstrated high sensitivity and specificity in removing misassigned transcripts. In real data applications, MisTIC effectively enhances cell type identification, reduces ambiguity in differential expression analysis, and improves the detection of cell-cell communication. Furthermore, RNA localization analysis based on data corrected by MisTIC revealed that, in T cells located near cancer-associated fibroblasts compared to those farther away, genes involved in T cell activation, inflammation, and cytotoxicity were depleted from cytoplasmic regions despite not being differentially expressed between the two T cell subsets. In conclusion, MisTIC is a powerful tool for correcting transcript misassignment in SRT data. It not only improves the accuracy of routine analyses but also enables novel investigations that provide deeper insights into the dynamics of gene expression.

bioinformatics↗

bioRxiv2025

Structural and biochemical studies of human PP2A(B55) holoenzyme and ENSA protein complex

The serine/threonine protein phosphatase 2A (B55) holoenzyme (PP2A(B55)), a key cell cycle regulator, is tightly regulated by ENSA, whose phosphorylation at Ser67 inhibits PP2A(B55). The structural basis for how pENSA orchestrates this inhibition to govern mitotic progression remains unclear. In this study, we disapproved the previous hypothesis of direct ENSA-PP2A A subunit interaction, and showed that the ENSA-PP2A(B55) interaction requires the PP2A holoenzyme. We then determined cryo-EM structure of the PP2A(B55)-ENSAS67D complex at 3.03 [A] resolution, which reveals four distinct regions of ENSA, each engaging specific sites on the B55 or C subunits. The structure reveals an extended, inhibited conformation of the PP2A(B55) holoenzyme, which contrasts sharply with the contracted, active state of apo PP2A(B55). pENSA binding induces the C subunit to shift away from the B55 subunit, with the pS67-containing motif inserting into the catalytic site of C subunit and locking the PP2A(B55) holoenzyme in an inactive state. Our work also revealed novel ligand-binding surfaces on B55, and expanded definition of the short linear motif (SliM). Overall, our work provided novel insights into PP2A regulation during cell cycle, and offers new avenues for the discovery of specific PP2A(B55) inhibitors and modulators.

biophysics↗

bioRxiv2025

A transition zone enriched WIF1⁺ basal cell subtype is associated with benign prostatic hyperplasia

ABSTRACT/SUMMARYThe cellular composition and disease susceptibilities of the distinct zones of the human prostate remain incompletely understood. Through extensive single-cell RNA sequencing (scRNA-seq) of benign regions from prostatectomy specimens, we identified a basal cell population expressing WIF1, VCAN, and NRG1, among other genes, that was significantly enriched in the transition zone (TZ). Benign prostatic hyperplasia (BPH) is a common condition that causes widespread morbidity and is nearly exclusively localized to the TZ. Analysis of previously published scRNA-seq datasets further confirmed that WIF1+ basal cells were significantly enriched in BPH compared to normal prostate. Pathway and cell-cell communication analyses revealed that this basal subtype is associated with programs related to cell proliferation, epithelial-mesenchymal transition (EMT), angiogenesis, and hormone response. Together, the molecular signature, zonal distribution, and pathway enrichment suggest that TZ-enriched WIF1+ basal cells may contribute to BPH pathogenesis by promoting epithelial and stromal remodeling.

cell biology↗

bioRxiv2025

Super-CUT&Tag: A Sensitive, Spatially Resolved Approach for Epigenomic Profiling in Tissue Sections

Spatial profiling of protein-DNA interactions is essential for understanding gene regulatory processes, but current methods suffer from low sensitivity, limited resolution, and complex workflows. Here we introduce Super-CUT&Tag (Solid-phase interface embedded with an Ultra-dense DNA barcode array to Profile Epigenomic landscapes for spatially Resolved CUT&Tag), a robust spatial epigenomic method that directly transfers in situ protein-DNA interactions from tissue sections onto a pre-fabricated spatially barcoded array. Incorporating 3D polyamidoamine (PAMAM) dendrimers markedly improves chromatin capture efficiency, achieving several-fold higher sensitivity compared to current spatial CUT&Tag approaches. Using Super-CUT&Tag, we profile H3K27ac spatial landscape and uncover previously hidden active-chromatin heterogeneity across mouse embryonic tissues and stages. Integration with spatial transcriptomics further reveals spatiotemporal enhancer regulation dynamics of Neurod2 during corticogenesis. By overcoming key bottlenecks in sensitivity, resolution and usability, Super-CUT&Tag establishes a powerful platform for spatial epigenomic profiling in complex tissues.

genetics↗

bioRxiv2025

AI recognizes convergent somatic hypermutation signatures to allow the discovery of variant-resilient broadly neutralizing antibodies

Elucidating the mechanisms by which broadly neutralizing antibodies (bnAbs) develop to confer durable immunity in humans is pivotal for the rational design of next-generation vaccines and therapeutics. VH3-53/3-66-encoded public antibodies are widely elicited in the population after the COVID-19 pandemic. Here, we isolated 15 VH3-53/3-66-encoded bnAbs from elite neutralizers who experienced sequential SARS-CoV-2 Omicron breakthrough infections. These bnAbs exhibit exceptional potency and breadth against circulating variants and, as members of a long-lived public antibody lineage, likely contribute to durable humoral protection. Genetic analysis of VH3-53/3-66-encoded bnAbs reveals a convergent somatic hypermutation pattern at seven positions in and around the CDR loops that emerges with repeated viral exposures and distinguishes them from non-bnAbs. Grafting these combined mutations onto clinically escaped antibodies broadens their breadth and restores neutralization activity. Structural analysis shows that these convergent mutations cooperatively remodel CDRs to bind and tolerate virus receptor-binding domain mutations. Using an AI-based antibody language model trained on data linking somatic hypermutations to binding affinity and neutralization, the model is able to identify mutational patterns predictive of breadth, enabling the discovery of a rare bnAb with protective activity against the latest variants from early-pandemic antibody repertoires. This demonstrates the feasibility of establishing an AI-empowered pipeline to identify mutation-tolerant bnAbs from early-pandemic antibody repertoires to fight fast-evolving newly introduced viruses.

immunology↗

bioRxiv2025

3' to 5' Translation of Circular RNAs?

For decades, the 5[->]3 direction of translation has been a central dogma of molecular biology(1). Here we present evidence that eukaryotic circular RNAs (circRNAs) can serve as templates for 3[->]5 backward translation (BT), yielding polypeptides with distinct sequence and structural features not found in canonical proteomes. Mining of more than 6,000 multi-omics datasets identified [~]1 million candidate BT proteins across eukaryotes, including 59,000 high-confidence human BT proteins supported by mass spectrometry. Genetic combinatorial experiments (KO and KI) and cell-free translation of synthetic circRNAs establish BT as a conserved mechanism. Loss-of-function studies of hs.circCAPN15 in human and dm.circROLS in Drosophila underscore the functional importance of BT-derived proteins. Our work challenges the long-standing unidirectional translation paradigm, expands the functional landscape of the genetic code, and reveals a hitherto hidden layer of proteomic complexity with broad implications for biology and therapeutics.

molecular biology↗

bioRxiv2025

Elucidation and engineering of arabinofuranosyltransferase to enable total de novo biosynthesis of paris saponins in yeast

Paris saponins (PSs) are structurally complex steroidal saponins that, due to their diverse glycosylation patterns, exhibit a range of significant pharmacological activities, including anti-tumor and antibacterial effects. However, incomplete characterization of the key enzymes responsible for glycosylation modifications has hindered their efficient heterologous biosynthesis. In this study, we reprogrammed the sugar donor specificity of a steroidal rhamnosyltransferase (UGT93M3) to enable the transfer of arabinofuranose (Araf). Through structural analysis, we identified key amino acid residues (368H/Q) that play an important role in determining Araf donor specificity. Guided by this insight, we successfully reconstructed the paris saponin I (PSI) biosynthetic pathway in Saccharomyces cerevisiae using engineered enzymes. To address challenges related to donor availability, we introduced UDP-sugar biosynthetic modules (UDP-Rha and UDP-Araf) into yeast. With this integrated platform, we were able to de novo produce a range of paris saponins, including diosgenin-3-O-glucosyl-(1[->]6)-glucoside (DGG), diosgenin-3-O-rhamnosyl(1[->]2) [glucosyl(1[->]6)]glucoside (DRGG) and paris saponin II. This work establishes a novel microbial platform for the sustainable production of paris saponins, particularly PSI, advancing the biosynthesis of steroidal glycosides and providing a potential strategy for the industrial-scale production of bioactive saponins. TeaserIn this study, we report the first complete de novo biosynthesis of four bioactive PSs, PSI, PSII, DGG and DRGG, in Saccharomyces cerevisiae from simple carbon source. Key advances include: (i) A single amino-acid switch (N368H) in the rhamnosyltransferase UGT93M3 endowed high-efficiency transfer of the rare five-membered arabinofuranose, solving a bottleneck in PS I biosynthesis; (ii) Elucidation of the molecular basis for sugar-donor specificity through AlphaFold3 docking and 300-ns molecular-dynamics simulations, revealing a histidine "latch" that stabilizes UDP-Araf in the catalytic pose; (iii) Construction of a 16-gene yeast chassis that integrates plant P450s, optimized glycosyltransferases, and de novo modules for UDP-rhamnose and UDP-arabinofuranose supply, achieving de novo microbial production of PS II, DGG, DRGG and PS I from glucose alone.

synthetic biology↗

bioRxiv2025

Unified Probabilistic Analysis of CyTOF: A Deep Generative Approach using CytoOne

Extracting meaningful biological signals from Cytometry by time-of-flight (CyTOF) data remains challenging due to heterogeneity, data characteristics, and the presence of various technical artifacts. Current analysis workflows typically rely on task-specific tools assembled into pipelines, which often make inconsistent distributional assumptions and fail to fully leverage the structure of the data. We present CytoOne, a unified probabilistic framework tailored for CyTOF data that integrates batch correction, differential analysis, and visualization within a single model. CytoOne is built upon a Bayesian hierarchical architecture inspired by Nouveau Variational Autoencoders (NVAE) and employs a novel quasi zero-inflated softplus-normal (QZIPN) likelihood to flexibly model the sparse and noisy nature of CyTOF measurements. We demonstrate via qualitative and quantitative evaluations that CytoOne effectively approximates both marginal and joint distributions of CyTOF data, removes batch-specific artifacts, enables fine-grained differential expression analysis, and facilitates interpretable embeddings for exploratory analysis.

bioinformatics↗