TSTie Shen

Tie Shen

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Tie Shen has a total of 66 co-authors. Here are the top co-authors by number of shared publications:

  • Lunxian Liu

    Key Laboratory of National Forestry and Grassland Administration on Biodiversity Conservation in Karst Mountainous Areas of Southwestern China, Engineering Research Center of Carbon Neutrality in Karst Areas, Ministry of Education, Key Laboratory of Environment Friendly Management on High Altitude Rhododendron Diseases and Pests, Institutions of Higher Learning in Guizhou Province, School of Life Science, Guizhou Normal University, Guiyang 550025, China

    6 shared publications
  • Zhengdong Zhang

    College of Computer Science

    6 shared publications
  • Xinglin Wang

    Key Laboratory of National Forestry and Grassland Administration on Biodiversity Conservation in Karst Mountainous Areas of Southwestern China, Engineering Research Center of Carbon Neutrality in Karst Areas, Ministry of Education, Key Laboratory of Environment Friendly Management on High Altitude Rhododendron Diseases and Pests, Institutions of Higher Learning in Guizhou Province, School of Life Science, Guizhou Normal University, Guiyang 550025, China

    5 shared publications
  • Xingpan Meng

    Key Laboratory of National Forestry and Grassland Administration on Biodiversity Conservation in Karst Mountainous Areas of Southwestern China, Engineering Research Center of Carbon Neutrality in Karst Areas, Ministry of Education, Key Laboratory of Environment Friendly Management on High Altitude Rhododendron Diseases and Pests, Institutions of Higher Learning in Guizhou Province, School of Life Science, Guizhou Normal University, Guiyang 550025, China

    4 shared publications
  • Ning Lv

    Guizhou Provincial Leading Talent Workstation for Protein Design and Biological Imaging Innovation, Key Laboratory of National Forestry and Grassland Administration on Biodiversity Conservation in Karst Mountainous Areas of Southwestern China, School of Life Science, Guizhou Normal University, Guiyang 550025, China

    3 shared publications
  • Qihang Zhou

    School of Cyber Sciences, Guizhou Normal University, Guiyang 550025, China

    3 shared publications

Latest Publications

  • MMU-DPI: Enhancing Generalization in Drug–Protein Interaction Prediction through Multimodal Learning and a Label Mix Strategy

    by Jiahao Wei, Tie Shen

    Accurate prediction of drug–protein interactions (DPIs) is crucial for accelerating the drug discovery process. However, the scarcity of experimentally validated interactions can limit the learning of transferable interaction patterns, particularly for previously unseen drugs and proteins. To address this fundamental challenge, we propose the MMU-DPI framework. A key component of this framework is a Label Mix strategy tailored to multimodal DPI prediction, which performs interpolation only in the label space while keeping the input modalities unchanged. This strategy provides stochastic soft-target regularization and improves generalization performance under reduced-data and independent external Cold-both evaluation settings. To effectively process and utilize multimodal data, MMU-DPI adopts a multimodal dual-branch architecture. The first branch uses a Message Passing Neural Network (MPNN) to extract structured representations from drug molecular graphs. It also uses a Convolutional Neural Network (CNN) to capture key biological and functional features from amino acid sequences. The second branch constructs a heterogeneous interaction graph and uses a Graph Attention Network (GAT) to learn deep contextual relationships between drugs and proteins. A learnable global fusion weight combines complementary branch logits to generate the final prediction for each drug–protein pair. Experimental results on multiple benchmark data sets demonstrate that MMU-DPI outperforms several state-of-the-art DPI prediction methods. Case studies further support the ability of MMU-DPI to identify potential DPIs. These results indicate that MMU-DPI can serve as a useful computational tool for drug discovery.

  • SQ-KFP: A Framework for Spatially Quantitative Metabolic Flux Analysis Enables Imaging the In Vivo Absolute Metabolic Enzymatic Reaction Rate

    by Lunxian Liu, Yuyun Xu, Xingpan Meng, Zhengdong Zhang, Yunfei Xu, Xinglin Wang, Ning Lv, Zhiwei Wen, Feiyang Wang, Wei Shao, et al.

    Alterations in metabolite concentrations often serve as direct drivers of phenotypic variation or disease onset. The changes in metabolite concentrations are directly dependent on the enzymatic reaction rates. However, spatially resolved imaging of the absolute metabolic enzymatic reaction rate (also known as metabolic flux) remains a critical unmet challenge in systems biology. We developed spatially quantitative kinetic flux profiling (SQ-KFP), a framework for quantitatively imaging the metabolic reaction rate, which has the potential for broad application to plant and animal tissues. This approach firstly realized the integration of quantitative flux analysis with mass spectrometry imaging, by generating isotope labeling data with coupled temporal and spatial resolution via image registration. SQ-KFP imaged the backward reaction rate of fumarase (vFUM) with spatial resolution as a proof-of-concept, revealing different reaction rate densities in petiole-lamina junctions and spatial decoupling between the metabolic reaction rate, metabolite concentration, and isotopic enrichment in leaves of Oxalis corymbosa and Medicago lupulina (validated by various detection methods). These results demonstrate that neither metabolite concentrations nor isotopic labeling values can substitute for reaction rates. Our method provides quantitative imaging of metabolic reaction rates and enables spatial flux analysis across diverse tissues and organisms.

  • Open Access

    Dynamic Accumulation and Transcriptional Regulation of Alkylamides in Developing Zanthoxylum planispinum var. Dintanensis Fruits

    by Hang Zhang, Ning Lv, Xinglin Wang, Huan Tian, Lunxian Liu, Tie Shen, Qingxiong Yang

    The accumulation dynamics and regulatory mechanisms of the alkylamides, the key pungent compounds in the fruits of Sichuan peppers, remain poorly understood. Using fruits of the Zanthoxylum planispinum var. dintanensis (Dintan) harvested at five key developmental stages, we comprehensively mapped the accumulation of numbering compounds and their underlying molecular drivers by integrating HPLC-based metabolite profiling and de novo transcriptomics. Total alkylamide content increased during development, with hydroxyl-α-sanshool (HαSS) being predominant. The contributions of hydroxyl-β-sanshool (HβSS) and hydroxyl-ε-sanshool (HεSS) increased in later stages. Cluster and correlation analyses identified 51 candidate genes strongly correlated (|r| ≥ 0.6) with HαSS accumulation, predominantly enriched in fatty acid and branched-chain amino acid metabolism pathways. The expression patterns of five stearoyl-CoA desaturase (SCD) genes, one long-chain acyl-CoA synthetase (ACSL/fadD), and one S-(hydroxymethyl)glutathione dehydrogenase/alcohol dehydrogenase (frmA) gene closely mirrored HαSS accumulation. In contrast, 3-oxoacyl-[acyl-carrier-protein] synthase II (fabF) and one β-ketoacyl-CoA synthase (KCS) gene exhibited a negative correlation. Accordingly, a positive regulatory network was constructed for HαSS accumulation. These findings revealed key candidate targets for deciphering the molecular basis of its unique flavor and for breeding high-pungency cultivars.

  • Open Access

    Spatial Heterogeneity of Metabolic Response to Drought Stress in Medicago lupulina L. Leaves

    by Xinglin Wang, Ning Lv, Yuyun Xu, Xingpan Meng, Yukun Jin, Hongbin Gao, Fei Li, Yin Yi, Lunxian Liu, Tie Shen

    Background: Drought stress is a primary environmental constraint limiting crop growth and productivity. Current drought-related plant research predominantly focuses on whole-leaf analyses, neglecting the spatial heterogeneity of metabolites within leaf tissues. Methods: This study combined transcriptomic and metabolomic approaches to investigate spatially distinct metabolic responses in marginal versus central regions of Medicago lupulina L. leaves under PEG-simulated drought. Results: Findings demonstrated that TCA cycle metabolites exhibited relative stability between leaf margins and centers under drought conditions, suggesting preserved core metabolic functionality in central tissues to sustain stress tolerance. Additionally, shikimic acid displayed a significantly reduced regional gradient in stressed tissues (PEG Margin vs. PEG Center) compared to controls. Phenylalanine, tryptophan, liquiritigenin, isoliquiritigenin, coproporphyrin III, and coproporphyrinogen III itself exhibited significantly increased internal gradient differences in stressed groups compared to control groups. The coordinated upregulation of key biosynthetic genes (e.g., TAT, AST, FNS II) in both the marginal and central regions of stressed leaves indicates a metabolic shift toward the biosynthesis of downstream defensive flavonoids. These metabolites and genes accumulated preferentially in margin regions of stressed leaves, indicative of localized activation of defense-associated metabolic pathways. Conclusions: This study reveals a spatially partitioned metabolic response to drought stress in M. lupulina leaves, where defensive metabolism is preferentially enhanced at the leaf margins while core metabolic homeostasis is maintained. These findings provide new spatial insights into plant drought acclimation and identify potential targets for improving crop resilience through the fine-tuning of local metabolism.


  • Open Access

    Molecular Mechanism of Cuscuta Haustorium Specialization Inferences from Transcriptome and Metabolome Analysis

    by Xingpan Meng, Ning Lv, Xinglin Wang, Qihang Zhou, Xu Zhang, Ximin Zhang, Zhengdong Zhang, Lunxian Liu, Tie Shen

    Background: Cuscuta australis R. Br. is a parasitic herbaceous plant that obtains nutrients by forming specialized structures called haustoria to invade host plants. Methods: In this study, we elucidated the differences in the gene expression regulation and metabolic characteristics between Cuscuta australis and Glycine max (Glycine max (L.) Merr. Var Williams) through comprehensive transcriptomic and metabolomic analyses. Results: The results demonstrated significant differences in the gene expression and metabolic features between the haustorium and the distal stem segments. The differentially expressed genes absorbed by Cuscuta australis from the soybean host influence amino acid metabolism, and the expression of the S-adenosylmethionine decarboxylase gene may affect the production of 5′-methylthioadenosine. A high expression of the chalcone synthase enzyme could lead to an increased daidzein content. Many Glycine max genes were also integrated into Cuscuta australis within the haustorium. Conclusions: This study systematically analyzed, for the first time, the significant differences in gene expression and metabolic characteristics between the haustoria and distal stem segments of Cuscuta. It also explored the nutrient absorption mechanisms of the host plant. Additionally, the research discovered that Cuscuta can absorb a substantial amount of host genes and adapt to its parasitic lifestyle through differential gene expression and metabolic changes. These findings provide important insights into the parasitic mechanisms of Cuscuta australis and lay the foundation for the development of effective control strategies.


  • Open Access

    Metabolome and transcriptomics analyses reveal quality differences between Camellia tachangensis F. C. Zhang and C. sinensis (L.) O. Kunzte

    by Yunfei Xu, Qihang Zhou, Xinglin Wang, Xingpan Meng, Zhengdong Zhang, Xu Zhang, Ximin Zhang, Suzhen Niu, Guiping Chen, Lunxian Liu, et al.

    Tea ranks among the top three most beloved non-alcoholic beverages worldwide and boasts significant economic and health benefits. In addition to Camellia sinensis (L.) O. Kuntze, and other Camellia plants in China are consumed by residents as tea drinks, which also have important economic value. The present study introduces one of the wild tea species, namely, Camellia tachangensis F. C. Zhang. We analyzed changes in metabolite abundance and gene expression patterns of C. tachangensis and C. sinensis using metabonomics and transcriptomics. We found 1056 metabolites, including 256 differential metabolites (67 upregulated and 189 downregulated). Additionally, transcriptome analysis revealed 8049 differentially expressed genes, with 4418 upregulated and 3631 downregulated genes. C. sinensis boasts a notable abundance of Amino acids, which can be attributed to its specific genetic makeup. In Theanine and Caffeine metabolic pathways, the levels of the majority of amino acids and caffeine tend to decrease. In Flavonoid biosynthesis, the levels of the Flavanone Fustin and Epicatechin are higher in C. tachangensis, while Epigallocatechin and Gallocatechin levels are higher in C. sinensis. This indicates that the metabolic components of C. sinensis and C. tachangensis are not identical, which may result in a unique flavor.


  • Open Access

    Predicting drug–protein interactions by preserving the graph information of multi source data

    by Jiahao Wei, Linzhang Lu, Tie Shen

    Examining potential drug–target interactions (DTIs) is a pivotal component of drug discovery and repurposing. Recently, there has been a significant rise in the use of computational techniques to predict DTIs. Nevertheless, previous investigations have predominantly concentrated on assessing either the connections between nodes or the consistency of the network’s topological structure in isolation. Such one-sided approaches could severely hinder the accuracy of DTI predictions. In this study, we propose a novel method called TTGCN, which combines heterogeneous graph convolutional neural networks (GCN) and graph attention networks (GAT) to address the task of DTI prediction. TTGCN employs a two-tiered feature learning strategy, utilizing GAT and residual GCN (R-GCN) to extract drug and target embeddings from the diverse network, respectively. These drug and target embeddings are then fused through a mean-pooling layer. Finally, we employ an inductive matrix completion technique to forecast DTIs while preserving the network’s node connectivity and topological structure. Our approach demonstrates superior performance in terms of area under the curve and area under the precision–recall curve in experimental comparisons, highlighting its significant advantages in predicting DTIs. Furthermore, case studies provide additional evidence of its ability to identify potential DTIs.


  • Open Access

    Metabolomic Profiling Using UPLC‒MS/MS Revealed Differential Changes in Metabolites of Three Tea Plant Varieties and Their Potential Antioxidant Activities

    by Lunxian Liu, Xinglin Wang, Qiuying Li, Han Dai, Qihang Zhou, Xingpan Meng, Zhongting Chen, Ximin Zhang, Zhengdong Zhang, Yingliang Liu, et al.

    Tea plants, as essential health beverages, are widely recognized for their rich content of antioxidant metabolites. This study utilized metabolomics analysis techniques to compare the metabolic differences between Camellia tachangensis F. C. Zhang, Camellia gymnogyna Hung T. Chang, and Camellia sinensis (L.) O. Kunzte, aiming to reveal their potential antioxidant activities. By applying UPLC–MS/MS technology, we analyzed metabolites in the samples and identified 1056 metabolites. We utilized statistical methods such as orthogonal partial least squares discriminant analysis (OPLS‐DA) and principal component analysis to evaluate and compare the variations in the samples. An intersection analysis of metabolites from the three types of tea plants identified 29 major differential metabolites. Functional annotation and pathway analysis of these differential metabolites were conducted, ultimately discussing 24 antioxidant‐related metabolites, primarily belonging to flavonoids, phenolic acids, and other categories. Among the three types of tea, C. tachangensis has the highest content of procyanidin B4, which is the most abundant substance in this tea. In terms of the impact of this substance on tea, it is known to have antioxidant properties and contributes to the overall antioxidant capacity of tea. In both C. gymnogyna and C. sinensis, the top 5 antioxidant metabolites include substances such as apigenin‐6,8‐di‐C‐glucoside (vicenin‐2), epigallocatechin, and apigenin‐5‐O‐glucoside. These substances belong to the flavonoid class and contribute to the antioxidant properties of these teas. Beyond the flavonoids and phenolic acids, other classes of metabolites in tea plants also exhibited significant antioxidant properties. These different metabolites play a key role in the antioxidant function of tea, contributing to their potential health benefits. The metabolites provide crucial insights for exploring and developing high‐antioxidant products that occur naturally.


  • Open Access

    UPLC–MS/MS and Gene Expression Research to Distinguish the Colour Differences of Rhododendron liliiflorum H. Lév

    by Jin Dai, Xinglin Wang, Xingpan Meng, Xu Zhang, Qihang Zhou, Zhengdong Zhang, Ximin Zhang, Yin Yi, Lunxian Liu, Tie Shen

    Among ornamental plants, the colour of the petals is an important feature. However, the reason for the colour differences of Rhododendron liliiflorum remains unclear. To reveal the differences in the colour of R. liliiflorum, high-efficiency liquid chromatographic collar (UPLC–MS/MS) technology was used to study the yellow and white parts of R. liliiflorum. A total of 1187 metabolites were identified in R. liliiflorum petals, including 339 flavonoid metabolites. Seventy-eight types of flavonoids in these metabolites were found in the yellow and white parts of R. liliiflorum petals, along with 11 other significantly enriched substances. Combining gene expression-related data with differential metabolite data demonstrated effects of enrichment in the flavanonols (fustin), flavonols (epiafzelechin and afzelechin), and flavanones (pinocembrin) of flavonoid biosynthesis; glyccitin, 6″-O-malonylgenistin, and 6-hydroxydaidzein of isoflavonoid biosynthesis; and anthocyanin biosynthesis of malvidin-3-O-galactoside (primulin), delphinidin-3-O-rutinoside, cyanidin-3-O-glucoside (kuromanin), and cyanidin-3-O-rutinoside (keracyanin), which are potentially the contributing factors responsible for the differences in petal colour in R. liliiflorum. This study establishes a connection between the differential metabolites underlying the color differences in the petals of R. liliiflorum and the gene expression in R. liliiflorum. This will provide a foundation for subsequent research on the regulation of flower color in R. liliiflorum and have profound implications for horticultural applications of R. liliiflorum.


  • Open Access

    Predicting drug-protein interactions by preserving the node connectivity and topological structure based on the multi source data

    by Jiahao Wei, Linzhang Lu, Tie Shen

    Examining potential drug-target interactions (DTIs) is a pivotal component of drug discovery and repurposing.Recently, there has been a significant rise in the use of computational techniques to predict DTIs. Nevertheless, previous investigations have predominantly concentrated on assessing either the connections between nodes or the consistency of the network’s topological structure in isolation. Such one-sided approaches could severely hinder the accuracy of DTI predictions. In this study, we propose a novel method called TTGCN, which combines heterogeneous graph convolutional neural networks (GCN) and graph attention networks (GAT) to address the task of DTI prediction. TTGCN employs a two-tiered feature learning strategy, utilizing GAT and residual GCN (R-GCN) to extract drug and target embeddings from the diverse network, respectively. These drug and target embeddings are then fused through a mean-pooling layer. Finally, we employ an inductive matrix completion technique to forecast DTIs while preserving the network’s node connectivity and topological structure.Our approach demonstrates superior performance in terms of Area Under the Curve (AUC) and Area Under the Precision-Recall Curve (AUPR) in experimental comparisons, highlighting its significant advantages in predicting DTIs. Furthermore, case studies provide additional evidence of its ability to identify potential DTIs.