Biology subjects

Shang, X.

Publications and source records attributed to Shang, X..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

DisConST: Deciphering Spatial Domains Using Distribution-aware Contrastive Learning for Spatial Transcriptomics

Spatial transcriptomics (ST) is a cutting-edge technology that provides comprehensive insights into gene expression patterns from a spatial perspective. A key research focus within this field is spatial domain identification, which is essential for exploring tissue organization, biological development, and disease mechanisms. Although methods have been developed, they still face challenges in modeling the gene expression information together with the spatial locations, resulting in suboptimal accuracy. We introduce DisConST (Distribution-aware Contrastive Learning for Spatial Transcriptomics), a novel deep-learning method designed to improve spatial domain detection within spatial transcriptomics datasets. DisConST addresses key challenges, such as the high dropout rates and the complex integration of spatial and gene expression data, by incorporating contrastive learning strategies that are aware of the underlying data distributions. It employs the zero-inflated negative binomial (ZINB) distribution, along with graph contrastive learning, to generate more informative latent representations. These representations efficiently integrate spatial positions, transcriptomic profiles, and cell-type proportions within spots. We validated DisConST across diverse ST datasets of tissues, organs, and embryos from various sequencing platforms in both normal and disease states. Our results consistently demonstrated that DisConST achieves superior spatial domain recognition accuracy compared to existing state-of-the-art methods. Furthermore, our experiments highlighted the utility of DisConST in advancing research on tissue organization, embryonic development, and tumor immune microenvironment dissection. The source code for DisConST is freely available at https://github.com/Zhenpm/DisConST/.

bioinformatics↗

bioRxiv2025

SPINK1-COL18A1 Crosstalk Shapes Epigenome and Drives Cancer Stemness in Pancreatic Ductal Adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) is a devastating cancer with an increasing incidence and extremely dismal prognosis. Discovery and mechanistic understanding of the genetic and epigenetic drivers of PDAC is therefore of critical importance. Here, we uncover serine protease inhibitor Kazal type 1 (SPINK1) as a putative determinant of clinical progression and demonstrate that it can have a non-canonical function as a regulator of epigenomic states of PDAC cells and associated extensive changes in gene expression. We show that SPINK1 expression, which varies in PDAC, is associated with key aggressive phenotypic cancer states and is correlated with the expression of stemness markers both in vitro and in patient samples. Mechanistically, our results strongly suggest that SPINK1 acts through a new signaling axis, by interacting with COL18A1 in the Golgi apparatus, promoting endostatin release and eventually inducing extensive histone H3 modifications. These results reveal a new function of SPINK1 in PDAC and suggest a potential therapeutic route to combat PDAC aggressiveness by targeting the SPINK1-COL18A1-endostatin signaling.

cancer biology↗

bioRxiv2025

Effects of DNA Origami-Based Nanoagent Design on Apoptosis Induction in a Large 3D Spheroid Model

The extrinsic activation of programmed cell death by FasR/CD95 is a promising minimally invasive strategy for cancer treatment. This can be leveraged using high-precision nanoscale therapeutics: Utilizing DNA origami for precise Fas ligand (FasL) presentation resulted in over 100 times more potent apoptosis induction in single cells. However, treating large, solid tumors poses challenges for DNA origami-based therapeutics, including drug distribution and altered cellular behavior. Here, we addressed these challenges using a 3D spheroid model. First, we assessed DNA origamis ability to penetrate tumor tissue, finding that penetration is influenced by the DNA origami size rather its structural flexibility. Second, we evaluated the apoptosis induction efficacy by DNA origami-FasL nanoagents within the spheroid model. The most potent nanoagents were able to completely eradicate all cells in the spheroid. Results indicated that apoptosis induction depended strongly on FasL attachment strategy rather than DNA origami design. Notably, only a rigid neutravidin linker for FasL attachment, rather than a flexible dsDNA linker, halted spheroid growth and fully eradicated all cancer cells. This study offers critical insights into designing DNA-based therapeutics for complex cellular environments and significantly advances DNA origami nanotherapeutic development, highlighting the impact of nanoagent design on cell fate decisions.

bioengineering↗

bioRxiv2024

A Multi-Property Optimizing Generative Adversarial Network for de novo Antimicrobial Peptide Design

Antimicrobial peptides (AMPs) play a crucial role in developing novel antiinfective drugs due to their broad-spectrum antimicrobial activity and lower likelihood of causing bacterial resistance. However, laboratory synthesis of AMPs is tedious and time-consuming. Existing computational methods have limited capability in optimizing multiple desired properties simultaneously. Here, we propose a Multi-Property Optimizing Generative Adversarial Network (MPOGAN), a feedback-loop framework that iteratively learns from data with multiple desired properties. This approach enables de novo design of AMPs with potent antimicrobial activity, reduced cytotoxicity, and diversity. Through extensive computational tests, MPOGAN exhibits superior performance in optimizing multiple desired properties of generated AMPs. Ten of the most promising candidates are chemically synthesized, with nine showing potent antimicrobial activity against three bacterial strains and low cytotoxicity against eukaryotic cells. MPOGAN offers a powerful computational approach for effective multi-property optimization of AMPs, thereby advancing the field of AI-aided drug discovery.

bioinformatics↗

bioRxiv2024

Tunability of calcium dynamics by signaling inputs and cell-cell communication in pancreatic beta cells

In addition to frequently occurring stable all-or-none responses, live cells can display more complex response dynamics, e.g., oscillations in the activity/concentration of biomolecules. While the emergence and function of oscillatory dynamics have been heavily investigated, fewer efforts have been spent on whether cells can fine-tune different aspects of an oscillatory signal (e.g., peakwidth, duty cycle, and frequency) and whether this fine-tuning can allow oscillatory signals to convey different information. In this study, we investigate glucose-induced calcium (Ca2+) oscillation in MIN6 cells, finding that the spontaneous or induced changes in the phosphatidylinositol 4,5-bisphosphate (PIP2) level during Ca2+ oscillation can modulate the peakwidth of individual Ca2+ spikes. Additionally, using a combination of optogenetics and Ca2+ imaging, we demonstrate that variation of the widths and frequencies of Ca2+ spikes can exert strong influence on coupling between neighboring MIN6 cells.

systems biology↗

bioRxiv2024

High yield purification of an Isoleucine zipper modified CD95 Ligand with either biotin or DNA-oligomer binding domain for efficient Cell Apoptosis Induction

BackgroundCluster of differentiation 95 (CD95/Fas/Apo1) as part of the Tumor-necrosis factor (TNF) receptor family is a prototypic trigger of the extrinsic apoptotic pathway and its activation by the trimeric ligand CD95L is of high interest for anticancer therapy. However, CD95L, when presented in solution, exhibits a low efficiency to induce apoptosis signaling in human cells. ResultsHere, we design a recombinant CD95L exhibiting an isoleucine zipper (IZ) motif at the N-terminus for stabilization of the trimerized CD95L and demonstrate its high apoptosis induction efficiency. A cysteine amino acid fused behind the IZ is further used as a versatile coupling site for bionanotechnological applications or for the development of biomedical assays. A fast, cheap, and high-yield production of CD95L via the HEK293T secretory expression system is presented, along with CD95L affinity purification and functionalization. We verified the biological activity of the purified protein and identified a stabilized trimeric CD95L structure as the most potent inducer of apoptosis signaling. ConclusionsThe workflow and the findings reported here will streamline a wide array of future low- or high-throughput TNF-ligand screens, and their modification towards improving apoptosis induction efficiency and anticancer therapy.

cell biology↗

bioRxiv2024

Efficient integration of multiple spatial transcriptomics data for 3D domain detection, matching, and alignment with stMSA

Spatial omics (SO) is a powerful methodology that enables the study of genes, proteins, and other molecular features within the spatial context of tissue architecture. With the growing availability of SO datasets, researchers are eager to extract biological insights from larger datasets for a more comprehensive understanding. However, existing approaches focus on batch effect correction, often neglecting complex biological patterns in tissue slices, complicating feature integration and posing challenges when combining transcriptomics with other omics layers. Here, we introduce stMSA (SpaTial Multi-Slice/omics Analysis), a deep graph contrastive learning model that incorporates graph auto-encoder techniques. stMSA is specifically designed to produce batch-corrected representations while retaining the distinct spatial patterns within each slice, considering both intra- and inter-batch relationships during integration. Extensive evaluations show that stMSA outperforms state-of-the-art methods in distinguishing tissue structures across diverse slices, even when faced with varying experimental protocols and sequencing technologies. Furthermore, stMSA effectively deciphers complex developmental trajectories by integrating spatial proteomics and transcriptomics data, and excels in cross-slice matching and alignment for 3D tissue reconstruction.

bioinformatics↗

bioRxiv2024

Deep Learning-Based Classification of CRISPR Loci Using Repeat Sequences

With the widespread application of the CRISPR-Cas system in gene editing and related fields, the demand for detecting and classifying CRISPR-Cas systems in metagenomic data has continuously increased. The traditional classification of the CRISPR-Cas system mainly relies on identifying neighboring cas genes of repeats. However, in some cases where there is a lack of information about cas genes, such as in metagenomes and fragmented genome assemblies, traditional classification methods may become ineffective. Here, we introduce a deep learning-based method called CRISPRclassify-CNN-Att, which classifies CRISPR-Cas systems solely based on repeat sequences. CRISPRclassify-CNN-Att utilizes convolutional neural networks (CNNs) and self-attention mechanisms to extract features from repeat sequences. It employs a stacking strategy to handle sample imbalances across different subtypes and improves classification accuracy for subtypes with fewer samples through transfer learning. CRISPRclassify-CNN-Att demonstrates excellent performance in classifying multiple subtypes, particularly in subtypes with a larger number of samples. Although CRISPR loci classification primarily relies on cas genes, CRISPRclassify-CNN-Att offers a new approach as a significant complement to current methods. It can identify unclassified loci missed by traditional cas-based methods, breaking the limitations of traditional approaches, and simplifying the classification process. The proposed tool is freely accessible via https://github.com/Xingyu-Liao/CRISPRclassify-CNN-Att.

bioinformatics↗

bioRxiv2024

The effects of L-carnitine and fructose in improved Ham's F10 on sperm culture in idiopathic severe asthenospermia within 24h

Trial designTo study the effects of L-carnitine and fructose on semen parameters of severe asthenospermia patients by sperm culturing in vitro within 24h. We optimized the energy composition and antioxidant substances of sperm culture medium in vitro (based on Hams F10 culture medium) by orthogonal test for preparing high quality culture medium. Sperms of 60 patients with idiopathic severe asthenospermia were collected, and cultured in vitro within 24h, by Hams F10 culture medium added to different concentrations of L-carnitine and fructose and culture temperature, whose effects on sperm motility were observed to determine which is the most appropriate concentration and temperature. MethodsFor determining the appropriate concentration of L-carnitine and fructose and the suitable culture temperature in Hams F10 culture medium, the orthogonal experiments were carried out to optimize above three factors, which had great influence on sperm viability, survival rate, deformity rate and DNA fragmentation index (DFI). ResultsThe final concentration of L-carnitine and fructose was determined in terms of initial tests to assess the effects of different concentrations (4, 8, 12, and 16 mg/ml L-carnitine and 0.125, 0.250, 0.375, and 0.50 mg/ml fructose) on sperm viability and motility in culture. During the operation of processing and culturing sperms in vitro within 24h, orthogonal test showed that sperm viability was better at the final concentration of 8 mg/ml L-carnitine and 0.375 mg/mL fructose in improved Hams F10 culture medium at 36.5 . ConclusionIdiopathic severe asthenospermia sperm can be effectively improved by the modified Hams F10 culture medium of the final concentration of 8 mg/ml L- carnitine and 0.375 mg/mL fructose in at 36.5 within 24h, which has shown better culture effect and is superior to Hams F10 basic medium.

biochemistry↗

bioRxiv2024

Unraveling the Genetic Comorbidity Landscape of Alzheimer's Disease

Alzheimers disease (AD) has emerged as the most prevalent and complex neurodegenerative disorder among the elderly population. However, the genetic comorbidity etiology for AD remains poorly understood. In this study, we conducted pleiotropic analysis for 41 AD phenotypic comorbidities, identifying ten genetic comorbidities with 16 pleiotropy genes associated with AD. Through biological functional and network analysis, we elucidated the molecular and functional landscape of AD genetic comorbidities. Furthermore, leveraging the pleiotropic genes and reported biomarkers for AD genetic comorbidities, we identified 50 potential biomarkers for AD diagnosis. Our findings deepen the understanding of the occurrence of AD genetic comorbidities and provide new insights for the search for AD diagnostic markers. HighlightsThe present study has focused on the comorbidities associated with Alzheimers disease (AD) by constructing a landscape of these comorbidities at various levels, including diseases, genetics, and pathways. O_LIThe study findings reveal novel and significant pathways that contribute to the etiology of AD and its comorbidities. C_LIO_LIBy exploring pleiotropic genes and reported biomarkers of AD comorbidities, the study has identified several potential diagnostic biomarker candidates for AD. C_LI O_FIG O_LINKSMALLFIG WIDTH=192 HEIGHT=200 SRC="FIGDIR/small/583453v1_figG1.gif" ALT="Figure 1"> View larger version (33K): org.highwire.dtl.DTLVardef@2fc040org.highwire.dtl.DTLVardef@f6d801org.highwire.dtl.DTLVardef@1df65a4org.highwire.dtl.DTLVardef@18a307e_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphic abstract.C_FLOATNO Study pipeline. C_FIG

neuroscience↗

bioRxiv2024

Targeting GFPT2 to reinvigorate immunotherapy in EGFR-mutated NSCLC

In the evolving field of cancer immunotherapy, EGFR-mutated NSCLC presents a significant challenge, demonstrating marked innate resistance to established treatments. An effective method to counter this resistance remains elusive. Through comprehensive genetic and pharmacological analyses across various models, we have identified glutamine fructose-6-phosphate transaminase 2 (GFPT2) as a key facilitator of immune evasion in EGFR-mutated NSCLC. Mechanistically, under EGFR mutation condition, GFPT2 expression, which is typically low in normal tissues, is highly induced via EGFR/IRE1/Xbp1s signaling axis, leading to a significant increase in intracellular UDP-GlcNAc and consequently, an altered N-glycosylation profile. GFPT2 escalates the expression and glycosylation of PD-L1, PVR and CD276, bolstering their interactions with CD8+T cells, and amplifies CD73 glycosylation, thereby intensifying adenosine-mediated CD8+T cells suppression. These actions collectively reduce tumor cell vulnerability to CD8+T cell-mediated death. Moreover, GFPT2 regulates EGFR glycosylation, which consequentially modulates the EGFR-dependent secretion of CXCL10 and VEGF, thus impeding CD8+T cell infiltration within tumors. We further identified a GFPT2 isoform-specific inhibitor that potentiates PD-1 blockade therapy beyond that of existing strategy, corroborated by results in xenografts and patient-derived organoids. Together, these findings illuminate the promising therapeutic potential of GFPT2 as a metabolic checkpoint, offering an innovative approach to invigorate immunotherapy in NSCLC with EGFR mutations.

cancer biology↗

bioRxiv2024

BERTE：High-precision hierarchical classification of transposable elements by a transfer learning method with BERT pre-trained model and convolutional neural network

Transposable Elements (TEs) are abundant repeat sequences found in living organisms. They play a pivotal role in biological evolution and gene regulation and are intimately linked to human diseases. Existing TE classification tools can classify classes, orders, and superfamilies concurrently, but they often struggle to effectively extract sequence features. This limitation frequently results in subpar classification results, especially in hierarchical classification. To tackle this problem, we introduced BERTE, a tool for TE hierarchical classification. BERTE encoded TE sequences into distinctive features that consisted of both attentional and cumulative k-mer frequency information. By leveraging the multi-head self-attention mechanism of the pre-trained BERT model, BERTE transformed sequences into attentional features. Additionally, we calculated multiple k-mer frequency vectors and concatenate them to form cumulative features. Following feature extraction, a parallel Convolutional Neural Network (CNN) model was employed as an efficient sequence classifier, capitalizing on its capability for high-dimensional feature transformation. We evaluated BERTEs performance on filtered datasets collected from 12 eukaryotic databases. Experimental results demonstrated that BERTE could improve the F1-score at different levels by up to 21% compared to current state-of-the-art methods. Furthermore, the results indicated that not only could BERT better characterize TE sequences in feature extraction, but also that CNN was more efficient than other popular deep learning classifiers. In general, BERTE classifies TE sequences with greater precision. BERTE is available at https://github.com/yiqichen-2000/BERTE.

bioinformatics↗

bioRxiv2024

Comprehensive Hierarchical Classification of Transposable Elements based on Deep Learning

Transposable elements (TEs) are DNA sequences capable of translocating within a genome. They constitute a substantial portion of eukaryotic genomes and play significant roles in genome evolution and gene regulation. The correct classification of these repetitive elements is essential to investigate their potential impact on genomes. Despite the existence of several tools for TE classification, they often neglect the importance of simultaneously utilizing global and local information for TE-type identification, resulting in suboptimal performance. Furthermore, these tools are not user-friendly due to the complex installation processes and numerous dependencies. In this study, we introduced a novel framework, CREATE, which leverages the strengths of Convolutional and Recurrent Neural NEtworks, combined with Attention mechanisms, for efficient TE classification. Given the tree-like structure of TE groups, we separately trained nine models within the class hierarchy. Benchmarking experiments showed that CREATE significantly outperformed other TE classification tools. The source code and demo data for CREATE are available at https://github.com/yangqi-cs/CREATE. To facilitate TE annotation for researchers, we have developed a web platform, named WebDLTE, based on the CREATE framework. This platform employs GPU-accelerated pre-trained deep learning models for real-time TE classification and offers the most comprehensive collection of TEs for download. The web interface can be accessed at https://www.webdlte.nwpu.edu.cn.

bioinformatics↗

bioRxiv2024

A multi-scale feature fusion model based on biological knowledge graph and transformer-encoder for drug-drug interaction prediction

Drug-Drug Interaction (DDI) refers to the combined effects that occur when a patient takes multiple medications simultaneously or within the same period. This interaction can either enhance the therapeutic effects of the drugs or inhibit their efficacy, and in severe cases, it can even lead to adverse drug reactions (ADRs). Thus, it is crucial to identify potential DDIs, as this information is significant for both biological research and clinical medicine. However, most existing works only consider the information of individual drugs or focus on the local correlation between a few medical entities, thus overlooking the global performance of the entire human medical system and the potential synergistic effects of multi-scale information. Consequently, these limitations hinder the predictive ability of models. In this paper, we propose an innovative multi-scale feature fusion model called ALG-DDI, which can comprehensively incorporate attribute information, local biological information, and global semantic information. To achieve this, we first employ the Attribute Masking method to obtain the embedding vector of the molecular graph. Next, ALG-DDI leverages heterogeneous graphs to capture the local biological information between drugs and several highly related biological entities. The global semantic information is also learned from the medicine-oriented large knowledge graphs. Finally, we employ a transformer encoder to fuse the multi-scale drug representations and feed the resulting drug pair vector into a fully connected neural network for prediction. Experimental evaluations on datasets of varying sizes and different classification tasks demonstrate that ALG-DDI outperforms other state-of-the-art models.

bioinformatics↗

bioRxiv2023

Structural insights into specific recognition of PTENα/β-NTE by WDR5 in cancer progression

PTEN/{beta}, two variants of PTEN, play a key role in promoting tumor growth by interacting with WDR5 through their N-terminal extensions (NTEs). This interaction facilitates the recruitment of the SET1/MLL methyltransferase complex, resulting in histone H3K4 trimethylation and upregulation of oncogenes such as NOTCH3, which in turn promotes tumor growth. However, the molecular mechanism underlying this interaction has remained elusive. In this study, we determined the crystal structure of PTEN-NTE in complex with WDR5, which reveal that PTEN specifically binds to the WIN site of WDR5 through a novel binding motif specifically found in the NTE domain of PTEN/{beta}. Disruption of this interaction significantly impedes cell proliferation and tumor growth, highlighting the potential of the WIN site inhibitors of WDR5 as therapeutic agents for PTEN/{beta} associated cancers. These findings not only shed light on the important role of the PTEN/{beta}-WDR5 interaction in carcinogenesis, but also present a promising avenue for developing cancer treatments that target this pathway.

cancer biology↗

bioRxiv2023

CTLA-4 tail fusion enhances CAR-T anti-tumor immunity

Chimeric antigen receptor (CAR) T cells are powerful therapeutics; however, their efficacy is often hindered by critical hurdles. Here, utilizing the endocytic feature of the cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) cytoplasmic tail (CT), we reprogram CAR function and substantially enhance CAR-T efficacy in vivo. CAR-T cells with monomeric, duplex, or triplex CTLA-4 CTs (CCTs) fused to the C-terminus of CAR exhibit a progressive increase in cytotoxicity under repeated stimulation, accompanied by reduced activation and production of pro-inflammatory cytokines. Further characterization reveals that CARs with increasing CCT fusion show a progressively lower surface expression, regulated by their constant endocytosis, recycling and degradation under steady state. The molecular dynamics of reengineered CAR with CCT fusion results in reduced CAR-mediated trogocytosis, loss of tumor antigen, and improved CAR-T survival. CARs with either monomeric (CAR-1CCT) or duplex CCTs (CAR-2CCT) have superior anti-tumor efficacy in a relapsed leukemia model. Single-cell RNA sequencing and flow cytometry analysis reveal that CAR-2CCT cells retain a stronger central memory phenotype and exhibit increased persistence. These findings illuminate a unique strategy for engineering therapeutic T cells and improving CAR-T function through synthetic CCT fusion, which is orthogonal to other cell engineering techniques.

immunology↗

bioRxiv2023

RAMEN identifies effective indicators for severe COVID and Long COVID patients

The COVID-19 pandemic has significantly altered global socioeconomic structures and individual lives. Understanding the disease mechanisms and facilitating diagnosis requires comprehending the complex interplay among clinical factors like demographics, symptoms, comorbidities, treatments, lab results, complications, and other metrics, and their relation to outcomes such as disease severity and long term outcomes (e.g., post-COVID-19 condition/long COVID). Conventional correlational methods struggle with indirect and directional connections among these factors, while standard graphical methods like Bayesian networks are computationally demanding for extensive clinical variables. In response, we introduced RAMEN, a methodology that integrates Genetic Algorithms with random walks for efficient Bayesian network inference, designed to map the intricate relationships among clinical variables. Applying RAMEN to the Biobanque quebecoise de la COVID-19 (BQC19) dataset, we identified critical markers for long COVID and varying disease severity. The Bayesian Network, corroborated by existing literature and supported through multi-omics analyses, highlights significant clinical variables linked to COVID-19 outcomes. RAMENs ability to accurately map these connections contributes substantially to developing early and effective diagnostics for severe COVID-19 and long COVID.

bioinformatics↗

bioRxiv2022

Network-based hub biomarker discovery for glaucoma

Glaucoma is an optic neuropathy, and the leading cause of irreversible blindness worldwide. However, the early detection of glaucoma remains challenging as chronic forms of glaucoma remain largely asymptomatic until considerable irreversible visual field deficits have ensued. Thus, biomarkers that facilitate early diagnosis and treatment for patients with a high risk of progression are critical. Network medicine approaches can be useful in identifying key relationships and important biomolecules for complex diseases. In this paper, we identified several hub biomarkers/drug targets for the diagnosis, treatment and prognosis for glaucoma and explored their associations for glaucoma based on human disease-biomarker and disease-target-drug networks. These results were verified by text-mining and genomic/epidemiology data. We also predicted the new application of BMP1 and MMP9 to diagnose glaucoma and confirm the theory of hub biomarkers with multiple clinical applications. Further, relevant pivotal pathways (regulation of the multicellular organismal process, regulation of localisation, and cytoplasmic vesicle for biomarkers; signal transduction and developmental process for targets) for these hub biomolecules were discovered, which may be foundations for future biomarker and drug target prediction for glaucoma. In conclusion, based on complex networks, hub biomolecules, essential pathways, and close diseases were identified for glaucoma in diagnosis, treatment and prognosis.

bioinformatics↗