Biology subjects

Liu, X.

Publications and source records attributed to Liu, X..

At least 487 records · Page 27Linked to original sources

bioRxiv2025

Ferret model of bleomycin-induced lung injury shares features of human idiopathic pulmonary fibrosis

Idiopathic pulmonary fibrosis (IPF) is a debilitating lung disease with limited therapeutic options. The development of effective therapies has been hindered by the lack of models that recapitulate key features of human disease. Here we report a bleomycin-induced ferret PF model characterized by an irreversible decrease in pulmonary compliance and an increase of opacification, accompanied by "honeycomb cyst-like" structures and "proximalization" of distal lung epithelium. Cellular and molecular analysis by single-nucleus RNA sequencing revealed a significant shift in distal lung epithelium towards proximal epithelial phenotype. Importantly, a histopathological pattern of bronchiolization encompassing divergent atypical epithelial cells and KRT17+/TP63+/KRT5low "basaloid-like" cells was present in the distal fibrotic lung lesions. Trajectory analysis revealed AT2 cells transition through multiple cell-states in bleomycin injured ferret lungs, particularly AT2 to KRT8high/KRT7low/SOX4+ and eventually to KRT8high/KRT7high/SFN+/TP63+/KRT5low "basaloid-like" cells. Further, immunofluorescence analyses demonstrated KRT7 and KRT8 populations reside overlaying the ACTA2-positive myofibroblasts in fibrotic foci, implying their pro-fibrogenic activity similar to human IPF lungs. Collectively, our results provide evidence that bleomycin-induced lung injury in ferrets recapitulates pathophysiological, cellular, and molecular features of human IPF, suggesting that they may be a reliable model for understanding mechanisms of IPF pathogenesis and for testing therapeutic strategies for treatment of IPF. Take Home MessageBleomycin-induced acute lung injury in the ferret recapitulates pathophysiological, cellular, and molecular features of human IPF; thus the ferret may be a reliable species for studying mechanisms of IPF pathogenesis and testing therapeutic strategies.

pathology↗

bioRxiv2025

Single-Cell Metabolic Imaging Reveals Glycogen Driven-Adaptations in Endothelial Cells

Endothelial dysfunction (ED) is a defining feature of diabetes mellitus (DM) and a key contributor to many metabolic and cardiovascular diseases. Endothelial cells (ECs) are known to be highly glycolytic and primarily rely on glucose to meet their energy demands. However, the role of glycogen metabolism in ECs remains poorly characterized due to a lack of suitable tools. Here, we utilize stimulated Raman scattering (SRS) microscopy to investigate subcellular glycogen metabolism in live ECs under stress conditions associated with highly prevalent diabetes and diabetic complications. We demonstrate that ECs exposed to a diabetes-mimicking milieu- high glucose and tumor necrosis factor (TNF-)- divert excess glucose toward subcellular glycogen storage, and that this storage capacity is significantly enhanced by the inhibition of glycogen synthase kinase 3 (GSK3). Pulse-chase experiments uncover glycogen dynamics and reveal that glycogen is rapidly mobilized under glucose starvation, highlighting its role as an immediate energy reserve in ECs. We further extend the capabilities of SRS metabolic imaging to visualize glutamine and lactate metabolism for the first time, directly showcasing the reliance of ECs on these alternative carbon substrates during glucose deprivation. Our results indicate that ECs containing glycogen exhibit a reduced immediate metabolic demand for these gluconeogenic substrates in the absence of extracellular glucose. These findings suggest that glycogen may play a broader role beyond energy reserves in ECs by modulating stress-responsive metabolic adaptations and may offer potential therapeutic opportunities to address diabetes-induced ED and related cardiometabolic diseases.

cell biology↗

bioRxiv2025

The Mycobacterium tuberculosis ESX effector promotes pyroptosis-dependent pathogenicity and dissemination

O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=126 SRC="FIGDIR/small/653593v1_ufig1.gif" ALT="Figure 1"> View larger version (43K): org.highwire.dtl.DTLVardef@8cd5e5org.highwire.dtl.DTLVardef@2ec22forg.highwire.dtl.DTLVardef@1ceb773org.highwire.dtl.DTLVardef@6dfba4_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphic AbstractC_FLOATNO C_FIG Mycobacterium tuberculosis (Mtb) remains a major global health threat, partly due to the extensive cytotoxicity induced during infection. Although GSDMD-mediated excessive pyroptosis promotes pathogen dissemination and tissue damage in tuberculosis, the mechanism remains poorly understood. Here, we identify the effector EccB5, a component of the Mtb ESX-5 secretion system, as a key driver of pyroptosis and hyperinflammatory responses. EccB5 enhances Mtb virulence by inducing pyroptosis of macrophages, promoting bacterial dissemination and exacerbating lung pathology. Conditional knockdown EccB5 increases host cell viability. Mechanistically, EccB5 directly interacts with GSDMD, strengthens its association with caspase-1, and facilitates caspase-1-mediated cleavage of GSDMD both in vitro and in vivo. In summary, our findings uncover a precise mechanism by which Mtb modulates host responses and advances its pathogenicity.

microbiology↗

bioRxiv2025

Corn rust population genomics reveals a cryptic virulent group and adaptive effectorome

Understanding the population structure of pathogens and the genetic determinants driving virulence gains is crucial for managing epidemic plant diseases. Puccinia polysora, a giga-scale fungal pathogen causing southern corn rust, has posed significant threat to global food security recently. Traditionally, P. polysora was considered clonal with minimal genetic variation. However, our population genomic and transcriptomic studies conducted in China, the emerging epicentre of the disease, have challenged this view. By adopting variant analyses appropriate to the dikaryotic nature of the pathogen, we discovered an unexpectedly clear population structure with six distinct groups. A cryptic group exhibits high virulence, facilitated by group-specific variation, and diversification of effectors. Although the Chinese population of P. polysora is predominantly asexual, internuclear exchange on some chromosomes have introduced recombination signals. The comprehensive pan-effectorome analyses revealed substantial presence/absence variation and alternative splicing events on effectors, shaping a highly adaptive effector repertoire in P. polysora. In conclusion, our findings highlight the tandem mapping on exploring clear genetic structure of dikaryotic species and reported the emergence of a virulent group and an adaptive effectorome of P. polysora. Effective containment strategies must be flexible to counter the threats posed by the unexpectedly dynamic evolution of this pathogen.

molecular biology↗

bioRxiv2025

Automated MagNet Enrichment Unlocks Deep and Cost-Effective LC-MS Plasma Proteomics

Plasma is an ideal material for proteomics due to its diverse protein content reflecting physiological and pathological states, and its compatibility with minimally invasive sampling. Deep proteomic profiling of plasma is hindered by the dominance of high-abundant proteins that mask the detection of low-abundant proteins. To overcome this, we compared five plasma protein enrichment methods, MagNet, ENRICHplus, ENRICHiST, EasySep, and EXONET, against neat plasma using LC-MS proteomics. All five methods substantially increased protein identifications, with MagNet, ENRICHplus, EasySep, and EXONET yielding up to 4200 proteins per sample, over 7-fold more than neat plasma, using a 44-minute gradient on the Evosep One and data-independent acquisition on the timsTOF Pro 2. These methods enriched extracellular vesicle-associated proteins while effectively depleting high-abundant proteins. We further optimized the cost-effective MagNet protocol by increasing the plasma-to-bead ratio and automated the workflow, including Evotip loading, on the Biomek i5 liquid handler. Automated MagNet demonstrated high reproducibility and a remarkably low total cost of just a few dollars per sample. This streamlined enrichment strategy enables scalable, high-throughput LC-MS plasma proteomics, supporting biomarker discovery across large clinical cohorts.

systems biology↗

bioRxiv2025

Triple-N Dataset: Non-human Primate Neural Responses to Natural Scenes

Understanding the neural mechanisms of visual perception requires data that encompass both large-scale cortical activity and the fine-grained dynamics of individual neurons. While the Natural Scenes Dataset (NSD) has provided substantial insights into visual processing in humans (Allen et al., 2022), its reliance on functional magnetic resonance imaging (fMRI) limits the exploration of individual neuron contributions. To bridge this gap, we present a new dataset: the Triple-N (Non-human Primate Neural Responses to Natural Scenes) dataset, which extends the NSD framework to non-human primates, incorporating single-neuron activity and fMRI recorded from the inferotemporal (IT) cortex. Using Neuropixels probes, we recorded neural responses from five macaques as they passively viewed 1,000 shared NSD images. The dataset includes 59 sessions across 27 sub-regions, capturing over 30,000 visual responsive units. Many recordings were obtained from fMRI-defined category-selective regions, such as face-, body-, scene-, and color-selective areas. Our dataset enables in-depth exploration of neural responses at multiple levels - from population dynamics to single-neuron activity, providing new insights into various aspects of visual processing, including the heterogeneity of object selectivity within functional regions and the temporal dynamics of responses to natural images. Furthermore, the dataset enables joint cross-species analyses, by integrating the macaque neural recordings with human fMRI data, offering a framework for comparing and aligning visual representations across primate species. Overall, our dataset provides a valuable resource for advancing our understanding of visual perception, bridging the gap between large-scale neuroimaging and fine-grained electrophysiological signals, while also facilitating the development of computational models of the high-level visual system.

neuroscience↗

bioRxiv2025

Loss of ATG5 expression in a subset of human prostate cancers promotes tumor growth through accumulation of p62

Loss-of-function mutations in autophagy-related (ATG) genes are rare in cancer. However, we report herein that ATG5 is fully deleted in [~]14% of prostate cancers (PCa), rivaling that of well-established tumor suppressor genes. ATG5 expression was downregulated at both mRNA and protein levels and was associated with poor patient survival. The DU145 PCa cell line, isolated from a brain metastasis, is entirely deficient in ATG5; and while ATG5 reintroduction restored autophagy, it dramatically inhibited tumor growth in vivo and led to near complete consumption of the multifunctional autophagy receptor/signaling protein, p62. Deletion of SQSTM1 confirmed that p62 was essential for tumor growth; and Reverse Phase Protein Array analysis revealed that p62 protein was significantly increased in prostate tumors, despite a reduction in mRNA expression. Thus, ATG5 appears to function as a novel tumor suppressor in a subset of prostate tumors and does so, at least in part, through autophagic degradation of p62.

cancer biology↗

bioRxiv2025

Adaptive Asymmetric Adjustment in Affective Evaluation and Reciprocity for Altruistic Behaviors from Exogenous Uncertainty to Certainty

In direct reciprocity, environmentally imposed exogenous uncertainty frequently decouples benefactors altruistic intentions from final outcomes. How do cooperative bonds remain resilient against this inevitable "noise"? Adopting a comprehensive levels-of-analysis framework across eight experiments, we integrate interpersonal paradigms, computational modeling, multivariate fMRI decoding, and third-party social evaluations to uncover a robust "asymmetric adjustment" in beneficiarys affective evaluation and reciprocity under exogenous uncertainty, systematically delineating its cognitive algorithmic rules, neural representations, and adaptive functions in sustaining human cooperation. Specifically, beneficiarys reciprocity amplifies for positive deviations from expectations while buffering against reciprocal reductions from unintended shortfalls. Modeling reveals this computational resilience is governed by an asymmetric intrinsic prosocial bias, extending beyond traditional frameworks of outcome/intention-based evaluation, prediction-error-based social learning, or loss aversion. Supported by theory-of-mind system, this "asymmetric adjustment" is perceived as morally superior to other calculations, thereby serving a critical adaptive function to sustain cooperation amidst the volatility of an uncertain world.

animal behavior and cognition↗

bioRxiv2025

Discovery and CryoEM Structure of FPM13, a Periplasmic Metalloprotein Unique to Francisella

We report the cryoEM structure of the Francisella protein FTN_1118, identifying it as a novel 13 kDa periplasmic protein unique to the Francisella genus, which we now designate FPM13 (Francisella Periplasmic Metalloprotein, 13 kDa) based on its structural and biochemical properties. FPM13 was serendipitously identified during purification of Francisella type VI secretion system (T6SS) effector proteins, co-purifying with them. Its identity, initially unknown, was established using the novel cryoID method. The structure reveals a symmetrical, donut-shaped 18-mer with 9-fold dihedral symmetry, formed by two stacked nonamers head-to-head. It measures [~]8 nm both in height and in outer diameter, and has a 3.5 nm central channel. The complex features a double-layered wall with an inner {beta}-sheet core and an outer -helical shell. Each monomer adopts a compact fold comprising an N-terminus {beta}-strand, an -helix and two additional {beta} strands at the C-terminus. The assembly is stabilized by inter-ring loop interactions and hydrophobic and electrostatic contacts between neighboring subunits. Biochemical analyses, as shown by APEX-biotinylation and Triton X-114 phase partitioning, confirmed that FPM13 is a soluble periplasmic protein. ICP-MS demonstrated that FPM13 binds iron and copper. Deletion of FPM13 in Francisella novicida strains caused no growth defects in macrophages or mice but show increased copper sensitivity under iron-depleted conditions, suggesting a role for FPM13 in metal transport or detoxification.

microbiology↗

bioRxiv2025

Atomic Structure and Phospholipid Binding Properties of the Francisella Type VI Secretion System Effector Protein PdpC

Francisella tularensis, a gram-negative facultative intracellular pathogen, causes tularemia, a potentially fatal infection. The Francisella pathogenicity island (FPI) encodes a type VI secretion system (T6SS), critical for the bacteriums ability to replicate within host cells and cause disease. PdpC, an FPI-encoded T6SS effector protein, is essential for phagosomal escape, intracellular replication, and virulence in animals, yet its structure and function remain poorly understood. Here, we expressed PdpC recombinantly and determined the cryoEM structure of PdpC at 3.4 [A] resolution, revealing a monomeric 156 kDa protein with a seahorse-shaped architecture (80 [A] x 75 [A] x 120 [A]). PdpC comprises five domains: an N-terminal domain (NTD) forming a prominent lateral head lobe, a central body domain, a C-terminal tail domain, a small wedge domain at the groove between the body and tail, and an unmodeled "mouth" domain that completes the seahorse-like shape. Functional studies demonstrate that PdpC binds specific host cell phospholipids, including phosphatidylinositol-3-phosphate. These findings provide a foundation for elucidating PdpCs mechanism of action and developing strategies to prevent or treat tularemia.

microbiology↗

bioRxiv2025

Endothelial AGO1 Drives Vascular Inflammation and Atherosclerosis via a Non-Canonical Nuclear Mechanism

BACKGROUNDEndothelial cell (EC) dysfunction is a cause and consequence of vascular inflammation and lipid dysregulation in atherosclerosis, yet the molecular drivers linking EC dysfunction to systemic metabolic derangements remain incompletely understood. Moreover, whether inhibiting an endogenous gene in ECs can impact liver function, lipid profile, and the vascular inflammation in the context of atherosclerosis has not been demonstrated. We previously identified Argonaute 1 (AGO1), a component of the RNA-induced silencing complex, as a regulator of EC function in angiogenesis and obesity. However, the role of endothelial AGO1 in vascular inflammation and liver function in the context of hyperlipidemia and atherosclerosis is unknown. METHODSEC-conditional AGO1 knockout (EC-AGO1-KO) and wildtype mice were subjected to pro-atherosclerotic models induced by AAV9-PCSK9 coupled with a Western diet or partial carotid ligation. Metabolic and vascular phenotype and gene expression were analyzed. In human liver sinusoidal and aortic ECs, AGO1 was knocked down using antisense oligonucleotides (ASO), followed by assessment of inflammatory responses (qPCR, RNA-seq, ELISA, and monocyte adhesion assays). To identify the molecular mechanisms linking AGO1 and EC inflammation, Cut&Tag sequencing, chromatin immunoprecipitation, immunofluorescence, proximal ligation assay, and co-immunoprecipitation were performed. The therapeutic effect of AGO1 inhibition was assessed using ASO-delivered via lipid nanoparticle (LNP) for systemic distribution and monocyte membrane-coated nanoparticles (MoNP) to target the inflamed endothelium. RESULTSEC-AGO1-KO mice exhibited improved plasma lipid profiles, reduced hepatic steatosis, inflammation, and fibrosis, and decreased aortic atherosclerotic burden. AGO1 knockdown in ECs attenuated inflammatory responses. Mechanistically, AGO1 interacted with NF-{kappa}B p65 and promoted p65 nuclear translocation and the transcriptional activation of pro-inflammatory genes, including ICAM1 and THBS1. AGO1-ASO delivered through LNP or MoNP achieved the anti-inflammatory, anti-hyperlipidemic, and anti-atherosclerotic effects, recapitulating the phenotypes observed with EC-AGO1-KO. CONCLUSIONSEndothelial AGO1 promotes vascular inflammation and liver dysfunction in the context of hyperlipidemia and atherosclerosis, in part through a non-canonical nuclear action of AGO1 as an NF-{kappa}B coactivator. Inhibition of endothelial AGO1 provides the dual benefits of ameliorating lipid dysregulation and suppressing vascular inflammation. These results highlight EC-AGO1 as a possible therapeutic target for atherosclerosis and cardiometabolic diseases.

pathology↗

bioRxiv2025

Phased ERK function on muscle stem cell plasticity in axolotl

The precise regulation of muscle stem cell (MuSC) plasticity remains poorly understood. While studies have suggested that Pax7+ MuSCs in vertebrate limbs are committed solely to myogenic lineages during regeneration, we previously observed during axolotl tail regeneration, Pax7+ MuSCs contribute to the fibroblast and chondrocyte mesodermal lineages. However, the underlying mechanism is unclear. Here, we show that the duration of ERK signaling plays an important role in MuSCs cell fate. We observed that ERK exhibits sustained activation after tail amputation that is followed by subsequent inhibition. We found that ERK activation initiates MuSCs plasticity switch toward a trunk fibroblast fate by repressing PAX7 expression. However, subsequent inhibition of ERK activity is essential for the transition of trunk fibroblast to fin fibroblast. In addition, we found that the TGF-{beta}/SMAD2 cascade as a downstream mediator of ERK-driven regulation of MuSC plasticity towards fibroblast and chondrocyte during tail regeneration. Together, we uncovered a new phased stem cell regulatory mechanism linking injury-induced ERK signaling to stem cell plasticity regulation, offering new insights with potential implications for regenerative medicine.

developmental biology↗

bioRxiv2025

Binding of extracellular vesicles to stretched von Willebrand factor promotes platelet activation

Von Willebrand factor (vWF), promoting platelet aggregation in various diseases such as COVID-19, malaria and cancer, is a huge multimeric glycoprotein. This extraordinary size makes vWF a unique shear stress sensing molecule. Below a critical shear stress, vWF is in a globular conformation that prevents platelet binding. Above the critical shear stress, vWF is stretched into platelet accessible fibers. Although previous studies have suggested that leukocytes or cancer cells can bind to vWF fibers, acting forces and the likelihood of cell adhesion has remained largely unexplored. Here, we report that vWF is a size-selective protein that prefers to interact with objects smaller than 4 m in diameter. Consistently, tumor cell-derived extracellular vesicles (EVs) were able to interact with vWF in parallel to platelets. Although whole tumor cells under flow were unable to bind to vWF per se, binding of EVs and platelets along the vWF fiber promoted platelet aggregation, which in turn entrapped circulating tumor cells. In conclusion, our study highlights the shear-sensitive nature of vWF and its ability to bring EVs and platelets together to enhance coagulation. While EVs-vWF-platelet aggregates may serve as novel biomarkers, their therapeutic disruption may prevent hypercoagulation in disease.

cancer biology↗

bioRxiv2025

Proteomic profiling of primary cilia in the developing brain uncovers newregulators of cortical development

In the developing brain, the primary cilia of radial glial cells extend from the surface of the lateral ventricle, serving as the signaling hub to integrate environmental cues critical for brain formation. Dysfunctions in ciliary proteins contribute to a wide range of brain structural abnormalities in ciliopathies, yet identifying the bona fide ciliary protein components in the brain remains a significant challenge. Here, using proximity labeling and quantitative proteomics, we systematically charted proteins localized to the cilia of radial glial cells in the dorsal and ventral regions of the embryonic brain. From this dataset, we identified cohorts of new molecules intrinsic to the cilia at distinct regions of the developing brain. We validated the cilium localization of several components of the translation machinery and studied the mechanistic roles of ciliary candidates previously linked to brain malformation, including Marcks, a key regulator of radial glial polarity, and Ckap2l, a protein associated with Filippi syndrome. These results revealed previously unrecognized ciliary mechanisms that regulate brain development. Thus, our brain proteomic dataset provides a unique resource for understanding ciliary functions in brain development and the molecular etiology of developmental disorders.

developmental biology↗

bioRxiv2025

Increasing Stemness Drives Prostate Cancer Progression, Plasticity, Therapy Resistance and Poor Patient Survival

BackgroundCancer progression is often accompanied by dedifferentiation and acquisition of stem cell-like properties (stemness). In prostate cancer (PCa), lineage plasticity and therapy resistance remain major clinical challenges, yet a unified quantitative transcriptomic framework connecting stemness, androgen receptor (AR) signaling, castration resistance, and disease progression across the PCa continuum is lacking. MethodsWe integrated 87,339 transcriptomic profiles from 33 preclinical and clinical datasets spanning the PCa continuum from normal prostate and treatment-naive primary PCa (Pri-PCa) to PCa treated with neoadjuvant ADT (nADT) and metastatic castration-resistant PCa (mCRPC), with single-cell RNA-seq analyses encompassing 115,197 cells. Cancer stemness was quantified using a transcriptome-derived mRNA-based Stemness Index (mRNAsi; hereafter Stemness), and a 12-gene PCa-Stem signature was developed to capture PCa-specific stemness. Stemness, PCa-Stem, canonical AR activity (c_AR-A), castration-reprogrammed AR activity (cr_AR-A), RB1-loss, PTEN-loss, and MYC activity signatures were quantified across cohorts. Functional validation included MYC inhibition and representative PCa-Stem signature gene depletion in PCa models. Clinical prognostic significance was evaluated in independent patient cohorts. ResultsThe Stemness score and c_AR-A increased concordantly during early prostate tumorigenesis but diverged with PCa progression: as Gleason grade increased, c_AR-A declined while Stemness continually increased. mCRPC exhibited the highest Stemness and lowest c_AR-A, a pattern recapitulated in Pten/Rb1/Trp53-deficient mouse models. Global Stemness increased progressively across the PCa continuum, was enriched in aggressive PAM50-LumB and PCS1 subtypes, associated with proliferative and lineage plasticity programs, and predicted poor patient survival. The newly derived 12-gene PCa-Stem signature provided a PCa-specific molecular representation of Stemness and tracked disease progression and poor patient survival. Network analyses identified a coordinated mitotic regulatory program linking MYC activity, RB1-loss, cr_AR-A, and the PCa-Stem signature. Spatial and single-cell transcriptomic analyses localized the PCa-Stem program to lineage plasticity-related epithelial cells and demonstrated progressive expansion of PCa-Stem epithelial cells during PCa progression. Functional perturbation of representative PCa-Stem signature genes, as well as genetic and pharmacological MYC inhibition, consistently suppressed Stemness-associated phenotypes in diverse PCa models. ConclusionsCancer Stemness quantitatively captures PCa aggressiveness, lineage plasticity, treatment resistance, disease progression, and poor patient survival. cr_AR-A, RB1 loss, and MYC activation cooperate to reinforce the high-Stemness state and therapy resistance in mCRPC. Collectively, our work establishes a trajectory-integrated transcriptomic framework defining cancer Stemness as a quantifiable molecular and clinical determinant of PCa aggressiveness, lineage plasticity, disease progression, therapy resistance and patient survival.

cancer biology↗

bioRxiv2025

Systemic trafficking of mRNA lipid nanoparticle vaccine following intramuscular injection generates potent tissue-specific T cell response

The mRNA lipid nanoparticles (LNPs) represent a new generation of vaccine carriers designed to elicit potent immune responses against infectious diseases and cancer. Despite the clinical success and rapid advancements in mRNA LNP technologies, the trafficking patterns of LNPs after intramuscular (i.m.) administration and the subsequent tissue-specific immunological effects have not been systematically characterized. Here, we report that trafficking of mRNA LNPs to different organs following i.m. injection is crucial for the induction of tissue-specific immunity beyond systemic immune response, particularly in tissue-resident CD8+ T cell generation, which is important for localized defense. By fine-tuning the composition of mRNA LNPs, trafficking patterns to systemic organs can be modulated, which can alter the resulting tissue-specific immune response. Formulations with a greater ability to enter the bloodstream can preferentially localize and transfect cells in specific organs like the liver, elicit stronger tissue-specific CD8+ T cell immune responses, and achieve enhanced efficacy in a liver tumor model. These findings highlight the potential to tailor mRNA LNP compositions to modulate trafficking following i.m. injection, thereby providing novel strategies for designing tissue-specific vaccines. Such strategies are particularly valuable for organ-specific diseases like cancer and infectious diseases, where tissue targeting and long-lasting immunity are essential for therapeutic success.

bioengineering↗

bioRxiv2025

BioMiner: A Multi-modal System for Automated Mining of Protein-Ligand Bioactivity Data from Literature

Protein-ligand bioactivity data published in the literature are essential for drug discovery, yet manual curation struggles to keep pace with rapidly growing literature. Automated bioactivity extraction remains challenging because it requires not only interpreting biochemical semantics distributed across text, tables, and figures, but also reconstructing chemically exact ligand structures (e.g., Markush structures). To address this bottleneck, we introduce BO_SCPLOWIOC_SCPLOWMO_SCPLOWINERC_SCPLOW, a multi-modal extraction framework that explicitly separates bioactivity semantic interpretation from ligand structure construction. Within BO_SCPLOWIOC_SCPLOWMO_SCPLOWINERC_SCPLOW, bioactivity semantics are inferred through direct reasoning, while chemical structures are resolved via a chemical-structure-grounded visual semantic reasoning paradigm, in which multi-modal large language models operate on chemically grounded visual representations to infer inter-structure relationships, and exact molecular construction is delegated to domain chemistry tools. For rigorous evaluation and method development, we further establish BO_SCPLOWIOC_SCPLOWVO_SCPLOWISTAC_SCPLOW, a comprehensive benchmark comprising 16,457 bioactivity entries curated from 500 publications. BO_SCPLOWIOC_SCPLOWMO_SCPLOWINERC_SCPLOW validates its extraction ability and provides a quantitative baseline, achieving an F1 score of 0.32 for bioactivity triplets. BO_SCPLOWIOC_SCPLOWMO_SCPLOWINERC_SCPLOWs practical utility is demonstrated via three applications: (1) extracting 82,262 data from 11,683 papers to build a pre-training database that improves downstream models performance by 3.9%; (2) enabling a human-in-the-loop workflow that doubles the number of high-quality NLRP3 bioactivity data, helping 38.6% improvement over 28 QSAR models and identification of 16 hit candidates with novel scaffolds; and (3) accelerating protein-ligand complex bioactivity annotation, achieving a 5.59-fold speed increase and 5.75% accuracy improvement over manual workflows in PoseBusters dataset. BO_SCPLOWIOC_SCPLOWMO_SCPLOWINERC_SCPLOW and BO_SCPLOWIOC_SCPLOWVO_SCPLOWISTAC_SCPLOW provide a scalable extraction methodology and a rigorous benchmark, paving the way to unlock bioactivity data that previously required extensive human effort. All data and code are available at GitHub.

bioinformatics↗

bioRxiv2025

Deciphering the Sequence Basis and Application of Transcriptional Initiation Regulation in Plant Genomes Through Deep Learning

Transcription initiation is a critical regulatory step in plant gene expression, yet its sequence determinants remain largely elusive. Here we introduce GenoRetriever, an interpretable deep learning model that deciphers the sequence basis of transcriptional initiation regulation across plant genomes. Trained on STRIPE-seq data from 16 soybean tissues and six other crop species, GenoRetriever identifies 27 core sequence motifs that govern transcription start site (TSS) selection and usage. The model predicts TSS locations and usage levels with high accuracy, as validated by in silico motif insertions, saturation mutagenesis, and CRISPR-Cas9 promoter editing. It further reveals that 31.85% of natural variation between wild and domesticated soybean drives shifts in promoter motif usage during domestication, and uncovers lineage-specific motif effects between monocots and dicots. This interpretable model and its user-friendly web server for promoter analysis and design make GenoRetriever both a methodological innovation and practical tool for plant functional genomics and crop improvement.

plant biology↗