bioRxiv Science⌕ Search

Biology subjects

ZHANG, L.

Publications and source records attributed to ZHANG, L..

17 recordsLinked to original sources
bioRxiv2026

Pancreatic Duct Cells as a Potential Source for Human Islet Neogenesis: Insights from Imaging Mass Cytometry

The question of whether islet neogenesis occurs in adult humans has been a subject of long-standing debate. To explore the characteristics of islet endocrine cells associated with pancreatic ducts, we employed imaging mass cytometry to examine pancreatic tissues from individuals across different age groups, including those with prediabetes or type 2 diabetes (T2D). Our analysis revealed the presence of all five pancreatic islet endocrine cell types, along with two types of non-hormone-expressing endocrine cells, located within or immediately adjacent to the ducts. These cells were most abundant in infancy, with a gradual decline observed through adulthood. Notably, ductal {beta} cells predominated in infancy, whereas ductal cells became more prevalent in adulthood, and significantly increased in the group aged over 60 years. Obesity further increased the ductal {beta} cells in the subjects aged over 60 years. Under prediabetic and T2D conditions, an increase in all duct-related endocrine cells was observed. These findings indicate that ductal cells may serve as a reservoir for new pancreatic endocrine cells, offering potential insights into the promotion of endogenous {beta} cell regeneration in diabetic patients. Highlights{bigcirc} Characterization of various islet endocrine cell types related to ducts in human pancreas. {bigcirc}The insulin-positive cells are the dominant cells among all duct-related islet endocrine cell types during the infancy period, however, the glucagon-positive cells become the dominant cells in adulthood. {bigcirc}T2D, Obesity, and aging are involved in the increase in the number of duct-related endocrine cells.

developmental biology↗
bioRxiv2026

A spatial-temporal atlas of human islet pathophysiology identifies a size-dependent trajectory from compensation to decompensation

Type 2 diabetes (T2D) is characterized by progressive islet dysfunction, yet the transition from functional adaptation to failure remains poorly defined within the native tissue architecture. Using multiplex imaging mass cytometry, we systematically analyzed human pancreatic islets across a spectrum of non-diabetic (ND), prediabetic (PreD), and T2D donors, revealing that pathological remodeling is profoundly size-dependent. This remodeling is manifested as a coordinated evolution of subpopulation abundance, endocrine cell proportions, structural integrity, and protein expression profiles, revealing that islets of different sizes undergo divergent fates during disease progression. We identified a size-dependent vulnerability spectrum where medium-and large-sized islets (>100 m in diameter) serve as the primary histopathological correlates of glycemic failure (HbA1c), exhibiting early density loss and structural disintegration. In contrast, small islets (30-100 {micro}m in diameter) exhibited compensatory hormone upregulation during PreD. Notably, diverging from the death of the small paradigm in Type 1 Diabetes, our data reveal a significant expansion of extra-islet endocrine clusters (EECs) that initiates during the compensatory stage, effectively preceding the onset of overt hyperglycemia. Finally, t-SNE clustering reconstructed a continuous trajectory of islet remodeling, capturing the phenotypic evolution of islets from normoglycemia through compensation to clinical decompensation. This study provides a high-resolution atlas of islet pathophysiology, offering new insights into T2D progression. HighlightsO_LIIslet remodeling during T2D progression is critically dependent on islet size. C_LIO_LIMedium and large islets density negatively correlates with HbA1c. C_LIO_LISmall islets exhibit resilience to metabolic stress during T2D progression. C_LIO_LIEECs expansion suggests a process of -cell-biased neogenesis. C_LIO_LIt-SNE maps the islet trajectory from compensation to decompensation. C_LI

pathology↗
bioRxiv2025

Biologically Informative NA Deconvolution (BIND) excavates hidden features of the proteome from missing values in large-scale datasets

The fast-advancing mass spectrometry and related technologies have greatly extended the depth of coverage in large-scale proteomics studies, including single-cell applications. As sample numbers grow rapidly, it is often challenging to interpret the proteins with missing values that are often presented as "NA" (not available). It could be the evidence of no expression, low expression below the detection threshold, or false negative detection due to technical issues. Existing methods for missing values imputation, while generally useful, rarely consider the non-random NA values that inform biological significance. In the current study, we developed Biologically Informative NA Deconvolution (BIND) that applies an adaptive neighborhood-based modeling to deconvolve the nature of NAs as "biological" (low/no expression) or technical (experimental errors). Applying to multiple cell line datasets and human tissue extracellular vesicle datasets, BIND excavated the NAs that indicated "hallmark absence" of unique proteins. This led to improvements in protein-protein interaction analysis and the identification of novel disease biomarkers. To facilitate its public accessibility, we compiled BIND into a web server that features functional online operations and interactive visualizations. Furthermore, we demonstrated that the BIND server could deconvolve the NAs and improve the analyses of single-cell proteomics datasets. Overall, BIND delineates the biological significance of missing values rather than treating them as a burden, providing a critical perspective for understanding the complex proteome in various biological contexts. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=144 SRC="FIGDIR/small/660508v1_ufig1.gif" ALT="Figure 1"> View larger version (52K): org.highwire.dtl.DTLVardef@1a9d7c4org.highwire.dtl.DTLVardef@194b2d9org.highwire.dtl.DTLVardef@169d3bborg.highwire.dtl.DTLVardef@cba0de_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗
bioRxiv2025

RIPK3-dependent sequential recruitment of MLKL and RIPK1 drives PANoptotic cell death and chemokine production

PANoptosis, an immunogenic programmed cell death (PCD) modality, integrates features of pyroptosis, apoptosis, and necroptosis through assembly of the PANoptosome complex. Despite its conceptualization as a distinct PCD form, PANoptosis remains controversial due to insufficient characterization of its morphological hallmarks and molecular regulation. This study aimed to investigate the molecular mechanisms underlying the assembly of PANoptosomes in the PANoptotic pathway. We identified a novel receptor-interacting protein kinase 3 (RIPK3)-initiated PANoptotic pathway that functions without pattern recognition receptors (PRRs) and the ASC (apoptosis-associated speck-like protein containing a CARD) inflammasome. Using multimodal imaging and biochemical approaches, we identify unique morphological signatures distinguishing PANoptotic cells from canonical pyroptotic, apoptotic, or necroptotic counterparts. Mechanistically, RIPK3 forms round homopolymeric scaffolds--distinct from necroptotic amyloid-like fibers--to sequentially recruit MLKL and RIPK1, forming a dynamic RIPK3-MLKL-RIPK1-FADD-caspase-8 complex (RIPK3-PANoptosome). This platform coordinates concurrent activation of pyroptotic, apoptotic, and necroptotic effectors. Cross-regulatory interactions between these pathways establish a homeostatic system where perturbations bias death modality into a certain cell death type, altering the death process and outcomes. Functionally, PANoptotic cells orchestrate chemokine secretion through parallel kinase-dependent (RIPK3-MLKL) and kinase-independent (RIPK1-IKK-NF-{kappa}B) mechanisms, driving macrophage recruitment. Our findings resolve the molecular logic of PANoptosome assembly, redefine PANoptosis as a tunable PCD paradigm, and establish its role in immunomodulation, providing a framework for targeting inflammatory cell death in disease.

cell biology↗
bioRxiv2024

Chemically Induced Nuclear Pore Complex Protein Degradation via E3 Ligase TRIM21

Despite the exciting progress of the bifunctional degrader molecules, also known as proteolysis-targeting chimeras (PROTACs), the rapidly expanding field is still significantly hampered by the lacking of available E3 ligase ligands. Our research bridges this gap by uncovering a series of small-molecule ligands to the E3 ligase TRIM21 through DNA-Encoded Library (DEL) technology. We confirmed their interaction with TRIM21 using crystallography and demonstrated their anti-proliferative effects across various cancer cell types. Furthermore, proteomic studies identified that the mRNA Export Factor GLE1, and the Nuclear Pore Complex Protein NUP155, were significantly down-regulated upon TRIM21 ligand treatment. This degradation required TRIM21 and was ubiquitin-proteasome-dependent. More specifically, NUP155 was the primary target for the TRIM21 ligands, while GLE1 was considered a passenger target upon the initial degradation of NUP155. Using immunofluorescence techniques, we further demonstrated that the degradation of GLE1 and NUP155 proteins impaired the integrity of the nuclear envelope, leading to cell death. Highlighted by this research, a novel mode of action has been discovered for the TRIM21 E3 ligase ligand, acting as a monovalent degrader that triggers de novo interaction with functional complex proteins and induces their degradation.

cell biology↗
bioRxiv2024

Fast-Part: Fast and Accurate Data Partitioning for Biological Sequence Analysis

Developing effective machine learning models for classifications of biological sequences depends heavily on the quality of the training and test datasets split. Existing tools are either computationally expensive, unable to maintain the desired level of similarity between the training and test datasets, or unable to retain training-test ratio stratification. Here, we present Fast-Part, a fast and accurate sequence data partitioning tool that ensures strict homology separation between the training and test datasets and the best possible training: test stratification ratio, and at the same time, is computationally fast. Fast-Part demonstrates rapid and accurate partitioning performance across diverse protein sequence datasets and maintains strict partitioning compared to the existing tools. Fast-Part can handle massive datasets and maintain strict homology partitioning.

bioinformatics↗
bioRxiv2024

DeepAge: Harnessing Deep Neural Network for Epigenetic Age Estimation FromDNA Methylation Data of human blood samples

Accurate prediction of biological age from DNA methylation data is a critical endeavor in understanding the molecular mechanisms of aging and developing age-related disease interventions. Traditional epigenetic clocks rely on linear regression or basic machine learning models, which often fail to capture the complex, non-linear interactions within methylation data. This study introduces DeepAge, a novel deep learning framework utilizing Temporal Convolutional Networks (TCNs) to enhance the prediction of biological age from DNA methylation profiles using selected CpGs by a Dual-Correlation based apparoach. DeepAge leverages a sequence-based approach with dilated convolutions to effectively capture long-range dependencies between CpG sites, addressing the limitations of prior models by incorporating advanced network architectures including residual connections and dropout regularization. The dual correlation feature selection enhances our models predictive capabilities by identifying the most age-relevant CpG sites. Our model outperforms existing epigenetic clocks across multiple datasets, offering significant improvements in accuracy and providing deeper insights into the epigenetic determinants of aging. The proposed method not only sets a new standard in age estimation but also highlights the potential of deep learning in biologically relevant feature extraction and interpretation, contributing to the broader field of computational biology and precision medicine.

bioinformatics↗
bioRxiv2024

PathoLM: Identifying pathogenicity from the DNA sequence through the Genome Foundation Model

Pathogen identification is pivotal in diagnosing, treating, and preventing diseases, crucial for controlling infections and safeguarding public health. Traditional alignment-based methods, though widely used, are computationally intense and reliant on extensive reference databases, often failing to detect novel pathogens due to their low sensitivity and specificity. Similarly, conventional machine learning techniques, while promising, require large annotated datasets and extensive feature engineering and are prone to overfitting. Addressing these challenges, we introduce PathoLM, a cutting-edge pathogen language model optimized for the identification of pathogenicity in bacterial and viral sequences. Leveraging the strengths of pre-trained DNA models such as the Nucleotide Transformer, PathoLM requires minimal data for fine-tuning, thereby enhancing pathogen detection capabilities. It effectively captures a broader genomic context, significantly improving the identification of novel and divergent pathogens. We developed a comprehensive data set comprising approximately 30 species of viruses and bacteria, including ESKAPEE pathogens, seven notably virulent bacterial strains resistant to antibiotics. Additionally, we curated a species classification dataset centered specifically on the ESKAPEE group. In comparative assessments, PathoLM dramatically outperforms existing models like DciPatho, demonstrating robust zero-shot and few-shot capabilities. Furthermore, we expanded PathoLM-Sp for ESKAPEE species classification, where it showed superior performance compared to other advanced deep learning methods, despite the complexities of the task.

bioinformatics↗
bioRxiv2024

ProtAlign-ARG: Antibiotic Resistance Gene Characterization Integrating Protein Language Models and Alignment-Based Scoring

The evolution and spread of antibiotic resistance pose a global health challenge. Whole genome and metagenomic sequencing pose a promising approach to monitoring the spread, but typical alignment-based approaches for antibiotic resistance gene (ARG) detection are inherently limited in the ability to detect new variants. Large protein language models could present a powerful alternative but are limited by databases available for training. Here we introduce ProtAlign-ARG, a novel hybrid model combining a pre-trained protein language model and an alignment scoring-based model to expand the capacity for ARG detection from DNA sequencing data. ProtAlign-ARG learns from vast unannotated protein sequences, utilizing raw protein language model embeddings to improve the accuracy of ARG classification. In instances where the model lacks confidence, ProtAlign-ARG employs an alignment-based scoring method, incorporating bit scores and e-values to classify ARGs according to their corresponding classes of antibiotics. ProtAlign-ARG demonstrated remarkable accuracy in identifying and classifying ARGs, particularly excelling in recall compared to existing ARG identification and classification tools. We also extended ProtAlign-ARG to predict the functionality and mobility of ARGs, highlighting the models robustness in various predictive tasks. A comprehensive comparison of ProtAlign-ARG with both the alignment-based scoring model and the pre-trained protein language model demonstrated the superior performance of ProtAlign-ARG.

bioinformatics↗
bioRxiv2024

MetaCompare 2.0: Differential ranking of ecological and human health resistome risks

BackgroundWhile there is increasing recognition of numerous environmental contributions to the spread of antibiotic resistance, quantifying the relative contributions of various sources remains a fundamental challenge. Similarly, there is a need to differentiate acute human health risks corresponding to exposure to a given environment, versus broader ecological risk of evolution and spread of antibiotic resistance genes (ARGs) across microbial taxa. Recent studies have proposed various methods of harnessing the rich information housed by metagenomic data for achieving such aims. Here, we introduce MetaCompare 2.0, which improves upon the original MetaCompare pipeline by differentiating indicators of human health resistome risk (i.e., potential for human pathogens to acquire ARGs) from ecological resistome risk (i.e., overall mobility of ARGs across a given microbiome). ResultsTo demonstrate the sensitivity of the MetaCompare 2.0 pipeline, we analyzed publicly available metagenomes representing a broad array of environments, including wastewater, surface water, soil, sediment, and human gut. We also assessed the effect of sequence assembly methods on the risk scores. We further evaluated the robustness of the pipeline to sequencing depth, contig count, and metagenomic library coverage bias through comparative analysis of a range of subsamples extracted from a set of deeply sequenced wastewater metagenomes. The analysis utilizing samples from different environments demonstrated that MetaCompare 2.0 consistently produces lower risk scores for environments with little human influence and higher risk scores for human contaminated environments affected by pollution or other stressors. We found that the ranks of risk scores were not measurably affected by different assemblers employed. The Meta-Compare 2.0 risk scores were remarkably consistent despite varying sequencing depth, contig count, and coverage. ConclusionMetaCompare 2.0 successfully ranked a wide array of environments according to both human health and ecological resistome risks, with both scores being strongly impacted by anthropogenic stress. We packaged the improved pipeline into a publicly-available web service that provides an easy-to-use interface for computing resistome risk scores and visualizing results. The web service is available at http://metacompare.cs.vt.edu/

bioinformatics↗
bioRxiv2023

AI becomes a masterbrain scientist

Recent rapid and unexpected advancements in Artificial Intelligence (AI) have dramatically shocked the world. Large language models, such as ChatGPT and GPT-4, have demonstrated remarkable potential in aggregating and organizing knowledge while providing insights and solutions to specific questions. In this study, we further explore the possibility of GPT-4 taking on a central role in a complete closed-loop biological research, functioning as a masterbrain scientist responsible for generating hypotheses, designing experiments, analyzing results, and drawing conclusions. Our findings suggest that AI has the potential to evolve into a genuine scientist, and could lead an unprecedented revolution in the area of science.

bioinformatics↗
bioRxiv2021

mobileOG-db: a manually curated database of protein families mediating the life cycle of bacterial mobile genetic elements

Currently available databases of bacterial mobile genetic elements (MGEs) contain both "core" and accessory MGE functional modules, the latter of which are often only transiently associated with the element. The presence of these accessory genes, which are often close homologs to primarily immobile genes, limits the usability of these databases for MGE annotation. To overcome this limitation, we analysed 10,776,212 protein sequences derived from seven MGE databases to compile a comprehensive database of 6,140 manually curated protein families that are linked to the "life cycle" (integration, excision, replication/recombination/repair, transfer, and stability/defense) of all major classes of bacterial MGEs. We overlay experimental information where available to create a tiered annotation scheme of high-quality annotations and annotations inferred exclusively through bioinformatic evidence. We additionally provide an MGE-class label for each entry (e.g., plasmid, integrative element) derived from the source database, and assign a list of keywords to each entry to delineate different MGE functional modules and to facilitate annotation. The resulting database, mobileOG-db (for mobile orthologous groups), provides a simple and readily interpretable foundation for an array of MGE-centred analyses. mobileOG-db can be accessed at mobileogdb.flsi.cloud.vt.edu/, where users can browse and design, refine, and analyse custom subsets of the dynamic mobilome.

genomics↗
bioRxiv2021

DeepBioGen: Generalizing predictions to unseen sequencing profiles via visual data augmentation

Predictive models trained on sequencing profiles often fail to achieve expected performance when externally validated on unseen profiles. While many factors such as batch effects, small data sets, and technical errors contribute to the gap between source and unseen data distributions, it is a challenging problem to generalize the predictive models across studies without any prior knowledge of the unseen data distribution. Here, this study proposes DeepBioGen, a sequencing profile augmentation procedure that characterizes visual patterns of sequencing profiles, generates realistic profiles based on a deep generative model capturing the patterns, and generalizes the subsequent classifiers. DeepBioGen outperforms other methods in terms of enhancing the generalizability of the prediction models on unseen data. The generalized classifiers surpass the state-of-the-art method, evaluated on RNA sequencing tumor expression profiles for anti-PD1 therapy response prediction and WGS human gut microbiome profiles for type 2 diabetes diagnosis.

bioinformatics↗
bioRxiv2021

DeepGeni: Deep generalized interpretable autoencoder elucidates gut microbiota for better cancer immunotherapy

Recent studies revealed that gut microbiota modulates the response to cancer immunotherapy and fecal microbiota transplantation has clinical benefit in melanoma patients during the treatment. Understanding microbiota affecting individual response is crucial to advance precision oncology. However, it is challenging to identify the key microbial taxa with limited data as statistical and machine learning models often lose their generalizability. In this study, DeepGeni, a deep generalized interpretable autoencoder, is proposed to improve the generalizability and interpretability of microbiome profiles by augmenting data and by introducing interpretable links in the autoencoder. DeepGeni-based machine learning classifier outperforms state-of-the-art classifier in the microbiome-driven prediction of responsiveness of melanoma patients treated with immune checkpoint inhibitors. DeepGeni-based machine learning classifier outperforms state-of-the-art classifier in the microbiome-driven responsiveness prediction of melanoma patients treated with immune checkpoint inhibitors. Also, the interpretable links of DeepGeni elucidate the most informative microbiota associated with cancer immunotherapy response.

bioinformatics↗
bioRxiv2020

Synergistic Effects of the DRD2/3 Antagonist ONC201 and Radiation in Glioblastoma

BackgroundGlioblastoma (GBM) is the deadliest of all brain cancers in adults. The current standard-of-care is surgery followed by radiotherapy and temozolomide, leading to a median survival time of only 15 months. GBM are organized hierarchically with a small number of glioma-initiating cells, responsible for therapy resistance and tumor recurrence, suggesting that targeting glioma-initiating cells could improve treatment response. ONC201 is a first-in-class anti-tumor agent with clinical efficacy in some forms of high-grade gliomas. Here we test its efficacy against GBM in combination with radiation. MethodsUsing patient-derived GBM lines and mouse models of GBM we test the effects of radiation and ONC201 on GBM self - renewal in vitro and survival in vivo. A possible resistance mechanism is investigated using RNA-Sequencing. ResultsTreatment of GBM cells with ONC201 reduced self-renewal, clonogenicity and cell viability in vitro. ONC201 exhibited anti-tumor effects on radioresistant GBM cells indicated by reduced self-renewal in secondary and tertiary glioma spheres. Combined treatment with ONC201 and radiation prolonged survival in syngeneic and patient-derived orthotopic xenograft mouse models of GBM. Subsequent transcriptome analyses after combined treatment revealed shifts in gene expression signatures related to quiescent GBM populations, GBM plasticity, and GBM stem cells. ConclusionsOur findings suggest that combined treatment with the DRD2/3 antagonist ONC201 and radiation improves the efficacy of radiation against GBM in vitro and in vivo through suppression of GICs without increasing toxicity in mouse models of GBM. A clinical assessment of this novel combination therapy against GBM is further warranted. Key points- Combined treatment of ONC201 and radiation exhibit anti-tumor effects on cells from primary and recurrent GBM - Combined treatment significantly prolongs survival in vivo - Combined treatment potentially targets the quiescent GBM cell population Importance of the StudyThe survival rates for patients with GBM are unacceptably low and novel treatment approaches are needed. This study provides evidence that a combination of radiation and the dopamine receptor antagonist ONC201 significantly prolongs survival in mouse models of glioma.

neuroscience↗
bioRxiv2019

Self-reported sleep duration and daytime napping are associated with renal hyperfiltration and microalbuminuria in apparently healthy Chinese population

BackgroundSleep duration affects health in various way. The objective of this study was to investigate the relationship between sleep duration, daytime napping and kidney function in a middle-aged apparently healthy Chinese population.\n\nMethodsAccording to self-reported total sleep and daytime napping duration, 33,850 participants aged 38 to 90 years old from 8 regional centers were divided into subgroups. Height, weight, waistline, hipline, blood pressure, biochemical index, FBG, PBG, HbA1c, creatinine and urinary albumin-creatinine ratio (UACR) were measured and recorded in each subject. Microalbuminuria was defined as UACR>=30 mg/g, CKD was defined as eGFR<60 ml/min and hyperfiltration was defined as eGFR>=135 ml/min. Multiple logistic regressions were applied to investigate associations between sleep and kidney function.\n\nResultsCompared to participants with [7-8]h/day sleep, ORs of >9 h/day, (8, 9]h/day and <6h/day sleep for microalbuminuria were 1.317 (1.200-1.446, p<0.001), 1.215 (1.123-1.315, p<0.001) and 1.218 (0.967-1.534, p=0.094). eGFR levels were U-shaped associated with sleep duration among subjects with >=90ml/min eGFR, and N-shaped associated with sleep duration among subjects with <90ml/min eGFR. OR of >9h/day sleep for hyperfiltration was 1.400 (1.123-1.745, p=0.003) among eGFR>=90 ml/min participants. Daytime napping had a negative effect on renal health. Compared to participants did not have napping habit, the ORs of (0, 1]h/day, (1, 1.5]h/day and >1.5h/day daytime napping for microalbuminuria were 1.477 (1.370-1.591, p<0.001), 1.217 (1.056, 1.403, p=0.007) and 1.447 (1.242, 1.687, p<0.001).\n\nConclusionsTotal sleep duration are U-shaped associated with renal health outcomes. Daytime napping had a negative effect on renal health.

epidemiology↗