MSH-DTI: multi-graph convolution with self-supervised embedding and heterogeneous aggregation for drug-target interaction prediction.

BACKGROUND: The rise of network pharmacology has led to the widespread use of network-based computational methods in predicting drug target interaction (DTI). However, existing DTI prediction models typically rely on a limited amount of data to extract drug and target features, potentially affecting the comprehensiveness and robustness of features. In addition, although multiple networks are used for DTI prediction, the integration of heterogeneous information often involves simplistic aggregation and attention mechanisms, which may impose certain limitations. RESULTS: MSH-DTI, a deep learning model for predicting drug-target interactions, is proposed in this paper. The model uses self-supervised learning methods to obtain drug and target structure features. A Heterogeneous Interaction-enhanced Feature Fusion Module is designed for multi-graph construction, and the graph convolutional networks are used to extract node features. With the help of an attention mechanism, the model focuses on the important parts of different features for prediction. Experimental results show that the AUROC and AUPR of MSH-DTI are 0.9620 and 0.9605 respectively, outperforming other models on the DTINet dataset. CONCLUSION: The proposed MSH-DTI is a helpful tool to discover drug-target interactions, which is also validated through case studies in predicting new DTIs.

DAS-DDI: A dual-view framework with drug association and drug structure for drug-drug interaction prediction.

In drug development and clinical application, drug-drug interaction (DDI) prediction is crucial for patient safety and therapeutic efficacy. However, traditional methods for DDI prediction often overlook the structural features of drugs and the complex interrelationships between them, which affect the accuracy and interpretability of the model. In this paper, a novel dual-view DDI prediction framework, DAS-DDI is proposed. Firstly, a drug association network is constructed based on similarity information among drugs, which could provide rich context information for DDI prediction. Subsequently, a novel drug substructure extraction method is proposed, which could update the features of nodes and chemical bonds simultaneously, improving the comprehensiveness of the feature. Furthermore, an attention mechanism is employed to fuse multiple drug embeddings from different views dynamically, enhancing the discriminative ability of the model in handling multi-view data. Comparative experiments on three public datasets demonstrate the superiority of DAS-DDI compared with other state-of-the-art models under two scenarios.

Property-guided few-shot learning for molecular property prediction with dual-view encoder and relation graph learning network.

Molecular property prediction is an important task in drug discovery. However, experimental data for many drug molecules are limited, especially for novel molecular structures or rare diseases which affect the accuracy of many deep learning methods that rely on large training datasets. To this end, we propose PG-DERN, a novel few-shot learning model for molecular property prediction. A dual-view encoder is introduced to learn a meaningful molecular representation by integrating information from node and subgraph. Next, a relation graph learning module is proposed to construct a relation graph based on the similarity between molecules, which improves the efficiency of information propagation and the accuracy of property prediction. In addition, we use a MAML-based meta-learning strategy to learn well-initialized meta-parameters. In order to guide the tuning of meta-parameters, a property-guided feature augmentation module is designed to transfer information from similar properties to the novel property to improve the comprehensiveness of the feature representation of molecules with novel property. A series of comparative experiments on four benchmark datasets demonstrate that the proposed PG-DERN outperforms state-of-the-art methods.

Entrance and origin of the extracranial vertebral artery found on computed tomography angiography.

To investigated morphological variability of vertebral artery (VA) origin and its entrance level into cervical transverse foramina by computed tomography angiography (CTA). To retrospectively investigated CTA of 223 subjects (446 VA courses). Investigated were origin of the VA and its level of entrance into vertebral transverse foramen with notification of the sex and side of variation. The VA entered the C6 transverse process in 91.70% of specimens (409 out of 446 VA courses). Abnormal entrance of VA was observed in 8.30% of specimens (37 VA courses), with the level of entrance into the C3, C4, C5, or C7 transverse foramen at 0.22%, 2.47%, 4.71% and 0.90% respectively. Comparably, the overall variability of abnormal origin of VA was 1.57% (7 out of 466 VA courses), in which the left vertebral arteries all arose from aortic arch. The variation rate of vertebral entrance rose up to 50% in abnormal origin subgroup. When comparing subgroups of subjects with normal and abnormal origin, there was significance difference in the frequency of entrance variation in the level of transverse foramen (p < 0.001). Abnormal entrance and origin of VA were observed in 8.30% and 1.57% of VA courses, which can be accurately appeared by CTA. Regarding to the subgroups of abnormal origin, the frequency of entrance variation was significantly increased in the level of transverse foramen compared to that of normal origin.

A convenient process to fabricate gelatin modified porous PLLA materials with high hydrophilicity and strength.

PLLA porous materials with high porosity were prepared by a gradual precipitation method and further modified by using different concentrations of gelatin aqueous solutions. Therefore, porous materials with different contents of gelatin coating were obtained. The micro morphology, crystallization, thermal performance, hydrophilicity and mechanical properties of the materials were evaluated by scanning electron microscopy (SEM), differential scanning calorimetry (DSC), X-ray diffraction (XRD), thermogravimetric analysis (TGA), water uptake ability tests and compression tests. It was found that the modified materials were formed by the stacking of nanosheets. The materials can maintain more than 80% porosity, high water uptake abilities and fast water uptake rates after modification. The compressive moduli of the materials were significantly improved from the initial sample with a value of 0.57 MPa to 46.41 MPa with gelatin modification. Due to the high porosity of materials, interconnected pore structures, and good surface hydrophilicity, the materials were expected to be widely used in the field of tissue engineering scaffolds, especially for bone substitutes, mainly due to their tunable and excellent mechanical properties.

Cerebral ischemic preconditioning reduces glutamate excitotoxicity by up-regulating the uptake activity of GLT-1 in rats.

Our previous study has shown that cerebral ischemic preconditioning (CIP) can up-regulate the expression of glial glutamate transporter-1 (GLT-1) during the induction of brain ischemic tolerance in rats. The present study was undertaken to further explore the uptake activity of GLT-1 in the process by observing the changes in the concentration of extracellular glutamate with cerebral microdialysis and high-performance liquid chromatography. The results showed that a significant pulse of glutamate concentration reached the peak value of sevenfold of the basal level after lethal ischemic insult, which was associated with delayed neuronal death in the CA1 hippocampus. When the rats were pretreated 2 days before the lethal ischemic insult with CIP which protected the pyramidal neurons against delayed neuronal death, the peak value of glutamate concentration decreased to 3.9 fold of the basal level. Furthermore, pre-administration of dihydrokainate, an inhibitor of GLT-1, prevented the protective effect of CIP on ischemia-induced CA1 cell death. At the same time, compared with the CIP + Ischemia group, the peak value of glutamate concentration significantly increased and reached sixfold of the basal level. These results indicate that CIP induced brain ischemic tolerance via up-regulating GLT-1 uptake activity for glutamate and then decreasing the excitotoxicity of glutamate.

Up-regulation of neuronal and astrocytic GLT-1 mRNA in the hippocampal CA1 subfield during the induction of brain ischemic tolerance in rats.

Our previous study showed that 3-min cerebral ischemic preconditioning (CIP) up-regulated the expression of glial glutamate transporter-1 (GLT-1) protein, which protects pyramidal neurons, allowing them to survive an 8-min ischemic insult that usually induces severe delayed neuronal death in the hippocampal CA1 subfield. In the present study, in situ hybridization and immunohistochemistry were used to observe whether GLT-1 mRNA is modulated and whether actrocytes and/or neurons express GLT-1 mRNA during the induction of brain ischemic tolerance. We observed that GLT-1 mRNA is expressed in neurons and astrocytes in the hippocampal CA1 subfield. The expression of GLT-1 mRNA was significantly down-regulated in both neurons and astrocytes after the 8-min lethal ischemic insult. CIP for 3 min increased the expression of GLT-1 mRNA in neurons and astrocytes, and induced the elongation of the astrocytic processes around pyramidal neurons. It may be concluded that CIP performed 2 days before lethal ischemic insult activated astrocytes, which resulted in an increased number of lengthened processes expressing high levels of GLT-1, which protected the neurons and allowed them to survive 8-min ischemic insult that is usually lethal to neurons in the hippocampal CA1 subfield.