Predicting Drug-Protein Interactions through Branch-Chain Mining and multi-dimensional attention network.

Identifying drug-protein interactions (DPIs) is crucial in drug discovery and repurposing. Computational methods for precise DPI identification can expedite development timelines and reduce expenses compared with conventional experimental methods. Lately, deep learning techniques have been employed for predicting DPIs, enhancing these processes. Nevertheless, the limitations observed in prior studies, where many extract features from complete drug and protein entities, overlooking the crucial theoretical foundation that pharmacological responses are often correlated with specific substructures, can lead to poor predictive performance. Furthermore, certain substructure-focused research confines its exploration to a solitary fragment category, such as a functional group. In this study, addressing these constraints, we present an end-to-end framework termed BCMMDA for predicting DPIs. The framework considers various substructure types, including branch chains, common substructures, and specific fragments. We designed a specific feature learning module by combining our proposed multi-dimensional attention mechanism with convolutional neural networks (CNNs). Deep CNNs assist in capturing the synergistic effects among these fragment sets, enabling the extraction of relevant features of drugs and proteins. Meanwhile, the multi-dimensional attention mechanism refines the relationship between drug and protein features by assigning attention vectors to each drug compound and amino acid. This mechanism empowers the model to further concentrate on pivotal substructures and elements, thereby improving its ability to identify essential interactions in DPI prediction. We evaluated the performance of BCMMDA on four well-known benchmark datasets. The results indicated that BCMMDA outperformed state-of-the-art baseline models, demonstrating significant improvement in performance.

Spatial-MGCN: a novel multi-view graph convolutional network for identifying spatial domains with attention mechanism.

MOTIVATION: Recent advances in spatial transcriptomics technologies have enabled gene expression profiles while preserving spatial context. Accurately identifying spatial domains is crucial for downstream analysis and it requires the effective integration of gene expression profiles and spatial information. While increasingly computational methods have been developed for spatial domain detection, most of them cannot adaptively learn the complex relationship between gene expression and spatial information, leading to sub-optimal performance. RESULTS: To overcome these challenges, we propose a novel deep learning method named Spatial-MGCN for identifying spatial domains, which is a Multi-view Graph Convolutional Network (GCN) with attention mechanism. We first construct two neighbor graphs using gene expression profiles and spatial information, respectively. Then, a multi-view GCN encoder is designed to extract unique embeddings from both the feature and spatial graphs, as well as their shared embeddings by combining both graphs. Finally, a zero-inflated negative binomial decoder is used to reconstruct the original expression matrix by capturing the global probability distribution of gene expression profiles. Moreover, Spatial-MGCN incorporates a spatial regularization constraint into the features learning to preserve spatial neighbor information in an end-to-end manner. The experimental results show that Spatial-MGCN outperforms state-of-the-art methods consistently in several tasks, including spatial clustering and trajectory inference.

Galectin-1 promotes angiogenesis and chondrogenesis during antler regeneration.

BACKGROUND: Deer antlers are the only known mammalian structure that undergoes full regeneration. In addition, it is peculiar because when growing, it contains vascularized cartilage. The differentiation of antler stem cells (ASCs) into chondrocytes while inducing endochondral extension of blood vessels is necessary to form antler vascularized cartilage. Therefore, antlers provide an unparalleled opportunity to investigate chondrogenesis, angiogenesis, and regenerative medicine. A study found that Galectin-1 (GAL-1), which can be used as a marker in some tumors, is highly expressed in ASCs. This intrigued us to investigate what role GAL-1 could play in antler regeneration. METHODS: We measured the expression level of GAL-1 in antler tissues and cells by immunohistochemistry, WB and QPCR. We constructed antlerogenic periosteal cells (APCs, one cell type of ASCs) with the GAL-1 gene knocked out (APCGAL-1-/-) using CRISPR-CAS9 gene editing system. The effect of GAL-1 on angiogenesis was determined by stimulating human umbilical vein endothelial cells (HUVECs) using APCGAL-1-/- conditioned medium or adding exogenous deer GAL-1 protein. The effect of APCGAL-1-/- on chondrogenic differentiation was evaluated compared with the APCs under micro-mass culture. The gene expression pattern of APCGAL-1-/- was analyzed by transcriptome sequencing. RESULTS: Immunohistochemistry revealed that GAL-1 was widely expressed in the antlerogenic periosteum (AP), pedicle periosteum (PP) and antler growth center. Western blot and qRT-PCR analysis using deer cell lines further supports this result. The proliferation, migration, and tube formation assays of human umbilical vein endothelial cells (HUVECs) showed that the proangiogenic activity of APCGAL-1-/- medium was significantly decreased (P < 0.05) compared with the APCs medium. The proangiogenic activity of deer GAL-1 protein was further confirmed by adding exogenous deer GAL-1 protein (P < 0.05). The chondrogenic differentiation ability of APCGAL-1-/- was impeded under micro-mass culture. The terms of GO and KEGG enrichment of the differentially expressed genes (DEGs) of APCGAL-1-/- showed that down-regulated expression of pathways associated with deer antler angiogenesis, osteogenesis and stem cell pluripotency, such as the PI3K-AKT signaling pathway, signaling pathways regulating pluripotency of stem cells and TGF-ß signaling pathway. CONCLUSIONS: Deer GAL-1, has strong angiogenic activity, is widely and highly expressed in deer antler. The APCs can induce angiogenesis by secreting GAL-1. The knockout of GAL-1 gene of APCs damaged its ability to induce angiogenesis and differentiate into chondrocytes. This ability is crucial to the formation of deer antler vascularized cartilage. Moreover, Deer antlers offer a unique model to explore explore how angiogenesis at high levels of GAL-1 expression can be elegantly regulated without becoming cancerous.

scGCL: an imputation method for scRNA-seq data based on graph contrastive learning.

MOTIVATION: Single-cell RNA-sequencing (scRNA-seq) is widely used to reveal cellular heterogeneity, complex disease mechanisms and cell differentiation processes. Due to high sparsity and complex gene expression patterns, scRNA-seq data present a large number of dropout events, affecting downstream tasks such as cell clustering and pseudo-time analysis. Restoring the expression levels of genes is essential for reducing technical noise and facilitating downstream analysis. However, existing scRNA-seq data imputation methods ignore the topological structure information of scRNA-seq data and cannot comprehensively utilize the relationships between cells. RESULTS: Here, we propose a single-cell Graph Contrastive Learning method for scRNA-seq data imputation, named scGCL, which integrates graph contrastive learning and Zero-inflated Negative Binomial (ZINB) distribution to estimate dropout values. scGCL summarizes global and local semantic information through contrastive learning and selects positive samples to enhance the representation of target nodes. To capture the global probability distribution, scGCL introduces an autoencoder based on the ZINB distribution, which reconstructs the scRNA-seq data based on the prior distribution. Through extensive experiments, we verify that scGCL outperforms existing state-of-the-art imputation methods in clustering performance and gene imputation on 14 scRNA-seq datasets. Further, we find that scGCL can enhance the expression patterns of specific genes in Alzheimer's disease datasets. AVAILABILITY AND IMPLEMENTATION: The code and data of scGCL are available on Github: https://github.com/zehaoxiong123/scGCL. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Evaluation of the tumoricidal efficacy of adoptive cell transfer using hepatocellular carcinoma-derived organoids.

Background: Tumor-derived organoid, namely tumoroid, can realistically retain the clinicopathologic features of original tumors even after long-term in vitro expansion. Here we develop this production methodology derived from hepatocellular carcinoma primary samples and generate a platform to evaluate the tumoricidal efficacy of autologous adoptive cell transfer including tumor infiltrating lymphocytes and peripheral blood lymphocytes. Methods: Haematoxylin and eosin together with immunohistochemistry staining were employed to ascertain the morphologic and histological features of tumoroids and original tumors. Tumor killing ability of T cells was detected by lactate dehydrogenase assay and propidium iodide staining. In tumoroid xenograft mouse model, tumor volumes were measured and T cell functions were examined by flow cytometry technique. Results: Four tumoroids with characteristics of poor differentiation and mild fibrosis were successfully established from fourteen hepatocellular carcinoma samples. More robust antitumor potential and hyper-functional phenotype of all four tumor infiltrating lymphocytes were observed compared to matched peripheral blood lymphocytes in coculture system. In tumoroid xenograft mouse models, however, only one patient-derived tumor infiltrating lymphocytes with the highest antitumor activity can bestow efficient tumor eradication. Conclusions: Hepatocellular carcinoma tumoroid-based models could represent invaluable resources for evaluating the tumoricidal efficacy of autologous adoptive cell transfer. Tumor infiltrating lymphocytes should be a promising and yet-to-be-developed regimen to treat hepatocellular carcinoma.

Constructing the in vitro culture system of the sika deer (cervus nippon) antler periosteal cell to detect its function on antler regeneration.

Periosteum is essential for bone regeneration and damage repair in mammals. Most species of deer family (Cervidae) develop two kinds of special periosteum, antler periosteum and pedicle periosteum, both supporting the complete regeneration of antler. Antler is the bone organ with the fastest growth rate in mammals. Along with the fast growth of antler, its external tissues such as blood vessels, nerves and the covering skin also grow rapidly. Currently, it is still unclear whether antler periosteum contributes to the fast growth of antler and how. It is also unclear why the regenerative capacity of antler periosteum is weaker than that of pedicle periosteum. In this study, the in vitro culture system for antler periosteal cells (AnPC) was constructed for the first time using the mid-beam antler periostea during antler fast-growth period. According to our results, the cultured AnPC expressed classical MSC markers, consistent with the pedicle periosteal stem cells (PPSC). However, the fluorescence intensities of the MSC markers on AnPC were significantly weaker than those on PPSC. In addition, AnPC showed much lower proliferation rates than PPSC. The proliferation rates of the AnPC also gradually decreased after successive passages, while the proliferation rates of the pedicle periosteal stem cells remained unchanged. These findings may partially explain the weaker regenerative capacity of antler periosteum. Further comparative global gene analysis revealed clearly the different gene expressed patterns between AnPC and PPSC. AnPC may mainly function on promoting angiogenesis, nerve growth and intramembrane bone formation during antler regeneration, whereas PPSC may primarily be involved in androgen signaling receptor pathway and PI3K-Akt signaling pathway and function on maintaining stem cell renewal.

Single-nucleotide polymorphisms of HLA and Polygonum multiflorum-induced liver injury in the Han Chinese population.

BACKGROUND: Polygonum multiflorum is one of the leading causes of herb-induced liver injury in China. HLA-B*35:01 is reported to be a potential biomarker of Polygonum multiflorum-induced liver injury (PM-DILI). However, little is known about the relationship between single-nucleotide polymorphisms (SNPs) and PM-DILI. AIM: To identify SNPs that indicate susceptibility to PM-DILI. METHODS: We conducted a systematic study enrolling 382 participants from four independent hospitals, including 73 PM-DILI patients, 118 patients with other drug-induced liver injury (other-DILI) and 191 healthy controls. Whole-exome sequencing was performed for 8 PM-DILI patients and 8 healthy controls who were randomly selected from the above subjects. Nineteen SNPs that showed high frequencies in the 8 PM-DILI patients were selected as candidate SNPs and then screened in 65 PM-DILI patients, 118 other-DILI patients and 183 healthy controls using the MassARRAY system. HLA-B high-resolution genotyping was performed for the 73 PM-DILI and 118 other-DILI patients. The Han-MHC database was selected as a population control for HLA-B analysis. P < 6.25 × 10-3 after Bonferroni correction was considered significant. RESULTS: The frequencies of rs111686806 in the HLA-A gene, rs1055348 in the HLA-B gene, and rs202047044 in the HLA-DRB1 gene were significantly higher in the PM-DILI group than in the control group [27.2% vs 11.6%, P = 1.72 × 10-5, odds ratio (OR) = 3.96, 95% confidence interval (CI): 2.21-7.14; 42.5% vs 8.6%, P = 1.72 × 10-19, OR = 13.62, 95%CI: 7.16-25.9; 22.9% vs 8.1%, P = 4.64 × 10-6, OR = 4.1, 95%CI: 2.25-7.47]. Only rs1055348 showed a significantly higher frequency in the PM-DILI group than in the other-DILI group (42.5% vs 13.6%, P = 1.84 × 10-10, OR = 10.06, 95%CI: 5.06-20.0), which suggested that it is a specific risk factor for PM-DILI. rs1055348 may become a tag for HLA-B*35:01 with 100% sensitivity and 97.7% specificity in the PM-DILI group and 100% sensitivity and 98.1% specificity in the other-DILI group. Furthermore, HLA-B*35:01 was confirmed to be associated with PM-DILI with a frequency of 41.1% in the PM-DILI group compared with 11.9% (P = 4.30 × 10-11, OR = 11.11, 95%CI: 5.57-22.19) in the other-DILI group and 2.7% (P = 6.22 × 10-166, OR = 62.62, 95%CI: 35.91-109.20) in the Han-MHC database. CONCLUSION: rs111686806, rs1055348, and rs202047044 are associated with PM-DILI, of which, rs1055348 is specific to PM-DILI. As a tag for HLA-B*35:01, rs1055348 may become an alternative predictive biomarker of PM-DILI.

A Review of Recent Developments and Progress in Computational Drug Repositioning.

Computational drug repositioning is an efficient approach towards discovering new indications for existing drugs. In recent years, with the accumulation of online health-related information and the extensive use of biomedical databases, computational drug repositioning approaches have achieved significant progress in drug discovery. In this review, we summarize recent advancements in drug repositioning. Firstly, we explicitly demonstrated the available data source information which is conducive to identifying novel indications. Furthermore, we provide a summary of the commonly used computing approaches. For each method, we briefly described techniques, case studies, and evaluation criteria. Finally, we discuss the limitations of the existing computing approaches.

Deficiency of canonical Wnt/ß-catenin signalling in hepatic dendritic cells triggers autoimmune hepatitis.

BACKGROUND & AIMS: Autoimmune hepatitis (AIH) is a chronic inflammatory liver disease manifested with the aberrant activation of hepatic dendritic cells (HDCs) and the subsequent breakdown of immune homeostasis. As an important player, HDC maintains immunological balance between tolerance to self-antigens versus destruction against pathogens in liver. However, the intracellular signalling networks that program HDC remain unclear. We have now found the role of canonical Wnt/ß-catenin signalling in HDCs. METHODS: Liver sections from AIH patients and healthy subjects were stained for the markers of Wnt/ß-catenin signalling. Concanavalin A (ConA) and HDC/Hepa1-6 vaccine-induced AIH mouse models were examined for liver injury, inflammation and immune cell functions by serum biochemistry, histology, quantitative reverse transcription polymerase chain reaction (qRT-PCR), enzyme-linked immunosorbent assay (ELISA) and flow cytometry analysis. Wnt/ß-catenin signalling expression was measured using immunoblot and qRT-PCR. RESULTS: Canonical Wnt/ß-catenin signalling in HDC is deficient in AIH patients and a mouse model, which coincides with the immunogenic function of HDCs. Furthermore, Wnt ligand engagement reactivates Wnt/ß-catenin signalling and recovers the immunoregulatory phenotype of HDCs, in turn alleviating the severity of AIH. Likewise, pharmacologic activation of Wnt/ß-catenin signalling attenuates AIH progression. CONCLUSIONS: We report here that the constitutively active canonical Wnt/ß-catenin signalling confers HDCs tolerogenicity under steady-state conditions. Deficiency of this pathway gives rise to T cell-mediated immune response and incidence of AIH. It may act as a new pathogenesis and treatment target for AIH.

Prediction of Disease Comorbidity Using HeteSim Scores based on Multiple Heterogeneous Networks.

BACKGROUND: Accumulating experimental studies have indicated that disease comorbidity causes additional pain to patients and leads to the failure of standard treatments compared to patients who have a single disease. Therefore, accurate prediction of potential comorbidity is essential to design more efficient treatment strategies. However, only a few disease comorbidities have been discovered in the clinic. OBJECTIVE: In this work, we propose PCHS, an effective computational method for predicting disease comorbidity. MATERIALS AND METHODS: We utilized the HeteSim measure to calculate the relatedness score for different disease pairs in the global heterogeneous network, which integrates six networks based on biological information, including disease-disease associations, drug-drug interactions, protein-protein interactions and associations among them. We built the prediction model using the Support Vector Machine (SVM) based on the HeteSim scores. RESULTS AND CONCLUSION: The results showed that PCHS performed significantly better than previous state-of-the-art approaches and achieved an AUC score of 0.90 in 10-fold cross-validation. Furthermore, some of our predictions have been verified in literatures, indicating the effectiveness of our method.