Recurrent neural network for trajectory tracking control of manipulator with unknown mass matrix.

Real-world robotic operations often face uncertainties that can impede accurate control of manipulators. This study proposes a recurrent neural network (RNN) combining kinematic and dynamic models to address this issue. Assuming an unknown mass matrix, the proposed method enables effective trajectory tracking for manipulators. In detail, a kinematic controller is designed to determine the desired joint acceleration for a given task with error feedback. Subsequently, integrated with the kinematics controller, the RNN is proposed to combine the robot's dynamic model and a mass matrix estimator. This integration allows the manipulator system to handle uncertainties and synchronously achieve trajectory tracking effectively. Theoretical analysis demonstrates the learning and control capabilities of the RNN. Simulative experiments conducted on a Franka Emika Panda manipulator, and comparisons validate the effectiveness and superiority of the proposed method.

Multiview representation learning for identification of novel cancer genes and their causative biological mechanisms.

Tumorigenesis arises from the dysfunction of cancer genes, leading to uncontrolled cell proliferation through various mechanisms. Establishing a complete cancer gene catalogue will make precision oncology possible. Although existing methods based on graph neural networks (GNN) are effective in identifying cancer genes, they fall short in effectively integrating data from multiple views and interpreting predictive outcomes. To address these shortcomings, an interpretable representation learning framework IMVRL-GCN is proposed to capture both shared and specific representations from multiview data, offering significant insights into the identification of cancer genes. Experimental results demonstrate that IMVRL-GCN outperforms state-of-the-art cancer gene identification methods and several baselines. Furthermore, IMVRL-GCN is employed to identify a total of 74 high-confidence novel cancer genes, and multiview data analysis highlights the pivotal roles of shared, mutation-specific, and structure-specific representations in discriminating distinctive cancer genes. Exploration of the mechanisms behind their discriminative capabilities suggests that shared representations are strongly associated with gene functions, while mutation-specific and structure-specific representations are linked to mutagenic propensity and functional synergy, respectively. Finally, our in-depth analyses of these candidates suggest potential insights for individualized treatments: afatinib could counteract many mutation-driven risks, and targeting interactions with cancer gene SRC is a reasonable strategy to mitigate interaction-induced risks for NR3C1, RXRA, HNF4A, and SP1.

Effects of transcranial direct current stimulation combined with Bosu ball training on the injury potential during drop landing in people with chronic ankle instability.

Purpose: To investigate the effects of transcranial direct current stimulation (tDCS) combined with Bosu ball training on the injury potential during drop landing in people with chronic ankle instability (CAI). Methods: A total of 40 participants with CAI were recruited and randomly divided into the tDCS + Bosu and Bosu groups. The people in the tDCS + Bosu group received intervention of tDCS combined with Bosu ball training, and those in the Bosu group received intervention of sham tDCS and Bosu ball training, for 6 weeks with three 20-min sessions per week. Before (week0) and after (week7) the intervention, all participants drop-landed on a trap-door device, with their affected limbs on a moveable platform, which could be flipped 24° inward and 15° forward to mimic an ankle inversion condition. The kinematic data were captured using a twelve-camera motion capture system. Two-way ANOVA with repeated measures was used to analyze data. Results: Significant group-by-intervention interactions were detected in the peak ankle inversion angular velocity (p = 0.047, η2 p = 0.118), the time to peak ankle inversion (p = 0.030, η2 p = 0.139), and the plantarflexion angle at the moment of peak ankle inversion (p = 0.014, η2 p = 0.173). Post hoc comparisons showed that compared with week0, the peak ankle inversion angular velocity and the plantarflexion angle at the moment of peak ankle inversion were reduced, the time to peak ankle inversion was advanced in both groups at week7, and the changes were greater in the tDCS + Bosu group compared to the Bosu group. And, a significant intervention main effect was detected in the peak ankle inversion angle in the two groups (p < 0.001, η2 p = 0.337). Conclusion: Compared with the Bosu ball training, the tDCS combined with Bosu ball training was more effective in reducing the injury potential during drop landing in people with CAI.

Modified double-pulley fixation provides better reduction of bone fragments and union compared to single-point fixation in bony Bankart lesions.

PURPOSE: The purpose of this study was to compare clinical scores and imaging outcomes of bony Bankart lesions that underwent single-point and modified double-pulley fixation after at least 2 years of follow-up. METHODS: Patients who underwent surgery to treat bony Bankart injuries were included and divided into groups A and B. A total of 69 patients were included (32 in group A and 37 in group B). Patients in group A underwent arthroscopic modified double-pulley fixation and patients in group B underwent arthroscopic single-point fixation. Three-dimensional computed tomography (3D-CT) was used to assess glenoid reduction one day after surgery. Postoperative bony union was assessed using 3D-CT and multiplanar reconstruction images 6 months after surgery. Constant-Murley, Rowe rating system, visual analogue scale and University of California at Los Angeles and American Shoulder and Elbow Surgeons scores were recorded before and after surgery. RESULTS: In terms of imaging measurements, there was no significant group difference in the preoperative size of the glenoid defect, the size of the bony fragment or the expected postoperative size of the glenoid defect. The sizes of the actual postoperative glenoid defects differed significantly between the groups (p = 0.027), as did the absolute difference between the expected and actual glenoid defect sizes (p < 0.001). At 6 months postoperatively, 50.0% of group A patients and 24.3% of group B patients exhibited complete bony union (p = 0.027); the rates of partial union were 37.5% and 56.8%, respectively. At the final follow-up, all clinical scores were significantly better than the preoperative scores (all p < 0.05), with no significant group differences (not significant). CONCLUSIONS: The use of the modified double-pulley technique with two anchors to treat bony Bankart injuries provides a better reduction of bone fragments than single-point fixation with two anchors and was associated with a higher rate of early bone union. LEVEL OF EVIDENCE: Level III.

Identification of key autophagy-related genes and pathways in spinal cord injury.

Spinal cord injury (SCI) can cause a range of functional impairments, and patients with SCI have limited potential for functional recovery. Previous studies have demonstrated that autophagy plays a role in the pathological process of SCI, but the specific mechanism of autophagy in this context remains unclear. Therefore, we explored the role of autophagy in SCI by identifying key autophagy-related genes and pathways. This study utilized the GSE132242 expression profile dataset, which consists of four control samples and four SCI samples; autophagy-related genes were sourced from GeneCards. R software was used to screen differentially expressed genes (DEGs) in the GSE132242 dataset, which were then intersected with autophagy-related genes to identify autophagy-related DEGs in SCI. Subsequently, the expression levels of these genes were confirmed and analyzed with gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG). A protein-protein interaction (PPI) analysis was conducted to identify interaction genes, and the resulting network was visualized with Cytoscape. The MCODE plug-in was used to build gene cluster modules, and the cytoHubba plug-in was applied to screen for hub genes. Finally, the GSE5296 dataset was used to verify the reliability of the hub genes. We screened 129 autophagy-related DEGs, including 126 up-regulated and 3 down-regulated genes. GO and KEGG pathway enrichment analysis showed that these 129 genes were mainly involved in the process of cell apoptosis, angiogenesis, IL-1 production, and inflammatory reactions, the TNF signaling pathway and the p53 signaling pathway. PPI identified 10 hub genes, including CCL2, TGFB1, PTGS2, FN1, HGF, MYC, IGF1, CD44, CXCR4, and SERPINEL1. The GSE5296 dataset revealed that the control group exhibited lower expression levels than the SCI group, although only CD44 and TGFB1 showed significant differences. This study identified 129 autophagy-related genes that might play a role in SCI. CD44 and TGFB1 were identified as potentially important genes in the autophagy process after SCI. These findings provide new targets for future research and offer new perspectives on the pathogenesis of SCI.

M2 Microglia-derived Exosomes Promote Spinal Cord Injury Recovery in Mice by Alleviating A1 Astrocyte Activation.

M2 microglia transplantation has previously demonstrated beneficial effects on spinal cord injury (SCI) by regulating neuroinflammation and enhancing neuronal survival. Exosomes (EXOs), secreted by almost all cell types, embody partial functions and properties of their parent cells. However, the effect of M2 microglia-derived EXOs (M2-EXOs) on SCI recovery and the underlying molecular mechanisms remain unclear. In this study, we isolated M2-EXOs and intravenously introduced them into mice with SCI. Considering the reciprocal communication between microglia and astroglia in both healthy and injured central nervous systems (CNSs), we subsequently focused on the influence of M2-EXOs on astrocyte phenotype regulation. Our findings indicated that M2-EXOs promoted neuron survival and axon preservation, reduced the lesion area, inhibited A1 astrocyte activation, and improved motor function recovery in SCI mice. Moreover, they inhibited the nuclear translocation of p65 and the activation of the NF-κB signalling pathway in A1 astrocytes. Therefore, our research suggests that M2-EXOs mitigate the activation of neurotoxic A1 astrocytes by inhibiting the NF-κB signalling pathway, thereby improving spinal tissue preservation and motor function recovery following SCI. This positions M2-EXOs as a promising therapeutic strategy for SCI.

[Comparison of intraoperative effects of computer navigation-assisted and simple arthroscopic reconstruction of posterior cruciate ligament tibial tunnel].

Objective: To compare the intraoperative effects of computer navigation-assisted versus simple arthroscopic reconstruction of posterior cruciate ligament (PCL) tibial tunnel. Methods: The clinical data of 73 patients with PCL tears who were admitted between June 2021 and June 2022 and met the selection criteria were retrospectively analysed, of whom 34 cases underwent PCL tibial tunnel reconstruction with navigation-assisted arthroscopy (navigation group) and 39 cases underwent PCL tibial tunnel reconstruction with arthroscopy alone (control group). There was no significant difference in baseline data between the two groups, including gender, age, body mass index, side of injury, time from injury to surgery, preoperative posterior drawer test, knee range of motion (ROM), Tegner score, Lysholm score, and International Knee Documentation Committee (IKDC) score between the two groups ( P>0.05). The perioperative indicators (operation time and number of guide wire drillings) were recorded and compared between the two groups. The angle between the graft and the tibial tunnel and the exit positions of the tibial tunnel in the coronal, sagittal, and transverse planes respectively were measured on MRI at 1 day after operation. The knee ROM, Tegner score, Lysholm score, and IKDC score were evaluated before operation and at last follow-up. Results: The operation time in the navigation group was shorter than that in the control group, and the number of intraoperative guide wire drillings was less than that in the control group, the differences were significant ( P<0.05). Patients in both groups were followed up 12-17 months, with an average of 12.8 months. There was no perioperative complications such as vascular and nerve damage, deep venous thrombosis and infection of lower extremity. During the follow-up, there was no re-injuries in either group and no revision was required. The results showed that there was no significant difference in the exit positions of the tibial tunnel in the coronal, sagittal, and transverse planes between the two groups ( P>0.05), but the angle between the graft and the tibial tunnel was significantly greater in the navigation group than in the control group ( P<0.05). At last follow-up, 30, 3, 1 and 0 cases were rated as negative, 1+, 2+, and 3+ of posterior drawer test in the navigation group and 33, 5, 1, and 0 cases in the control group, respectively, which significantly improved when compared with the preoperative values ( P<0.05), but there was no significant difference between the two groups ( P>0.05). At last follow-up, ROM, Tegner score, Lysholm score, and IKDC score of the knee joint significantly improved in both groups when compared with preoperative values ( P<0.05), but there was no significant difference in the difference in preoperative and postoperative indicators between the two groups ( P>0.05). Conclusion: Computer-navigated arthroscopic PCL tibial tunnel reconstruction can quickly and accurately prepare tunnels with good location and orientation, with postoperative functional scores comparable to arthroscopic PCL tibial tunnel reconstruction alone.

[Experimental study of M2 microglia transplantation promoting spinal cord injury repair in mice].

Objective: To investigate the effect of M2 microglia (M2-MG) transplantation on spinal cord injury (SCI) repair in mice. Methods: Primary MG were obtained from the cerebral cortex of 15 C57BL/6 mice born 2-3 days old by pancreatic enzyme digestion and identified by immunofluorescence staining of Iba1. Then the primary MG were co-cultured with interleukin 4 for 48 hours (experimental group) to induce into M2 phenotype and identified by immunofluorescence staining of Arginase 1 (Arg-1) and Iba1. The normal MG were harvested as control (control group). The dorsal root ganglion (DRG) of 5 C57BL/6 mice born 1 week old were co-cultured with M2-MG for 5 days to observe the axon length, the DRG alone was used as control. Forty-two 6-week-old female C57BL/6 mice were randomly divided into sham group ( n=6), SCI group ( n=18), and SCI+M2-MG group ( n=18). In sham group, only the laminae of T 10 level were removed; SCI group and SCI+M2-MG group underwent SCI modeling, and SCI+M2-MG group was simultaneously injected with M2-MG. The survival of mice in each group was observed after operation. At immediate (0), 3, 7, 14, 21, and 28 days after operation, the motor function of mice was evaluated by Basso Mouse Scale (BMS) score, and the gait was evaluated by footprint experiment at 28 days. The spinal cord tissue was taken after operation for immunofluorescence staining, in which glial fibrillary acidic protein (GFAP) staining at 7, 14, and 28 days was used to observe the injured area of the spinal cord, neuronal nuclei antigen staining at 28 days was used to observe the survival of neurons, and GFAP/C3 double staining at 7 and 14 days was used to observe the changes in the number of A1 astrocytes. Results: The purity of MG in vitro reached 90%, and the most of the cells were polarized into M2 phenotype identified by Arg-1 immunofluorescence staining. M2-MG promoted the axon growth when co-cultured with DRGs in vitro ( P<0.05). All groups of mice survived until the experiment was completed. The hind limb motor function of SCI group and SCI+M2-MG group gradually recovered over time. Among them, the SCI+M2-MG group had significantly higher BMS scores than the SCI group at 21 and 28 days ( P<0.05), and the dragging gait significantly improved at 28 days, but it did not reach the level of the sham group. Immunofluorescence staining showed that compared with the SCI group, the SCI+M2-MG group had a smaller injury area at 7, 14, and 28 days, an increase in neuronal survival at 28 days, and a decrease in the number of A1 astrocytes at 7 and 14 days, with significant differences ( P<0.05). Conclusion: M2-MG transplantation improves the motor function of the hind limbs of SCI mice by promoting neuron survival and axon regeneration. This neuroprotective effect is related to the inhibition of A1 astrocytes polarization.

Feature fusion network based on few-shot fine-grained classification.

The objective of few-shot fine-grained learning is to identify subclasses within a primary class using a limited number of labeled samples. However, many current methodologies rely on the metric of singular feature, which is either global or local. In fine-grained image classification tasks, where the inter-class distance is small and the intra-class distance is big, relying on a singular similarity measurement can lead to the omission of either inter-class or intra-class information. We delve into inter-class information through global measures and tap into intra-class information via local measures. In this study, we introduce the Feature Fusion Similarity Network (FFSNet). This model employs global measures to accentuate the differences between classes, while utilizing local measures to consolidate intra-class data. Such an approach enables the model to learn features characterized by enlarge inter-class distances and reduce intra-class distances, even with a limited dataset of fine-grained images. Consequently, this greatly enhances the model's generalization capabilities. Our experimental results demonstrated that the proposed paradigm stands its ground against state-of-the-art models across multiple established fine-grained image benchmark datasets.

HimGNN: a novel hierarchical molecular graph representation learning framework for property prediction.

Accurate prediction of molecular properties is an important topic in drug discovery. Recent works have developed various representation schemes for molecular structures to capture different chemical information in molecules. The atom and motif can be viewed as hierarchical molecular structures that are widely used for learning molecular representations to predict chemical properties. Previous works have attempted to exploit both atom and motif to address the problem of information loss in single representation learning for various tasks. To further fuse such hierarchical information, the correspondence between learned chemical features from different molecular structures should be considered. Herein, we propose a novel framework for molecular property prediction, called hierarchical molecular graph neural networks (HimGNN). HimGNN learns hierarchical topology representations by applying graph neural networks on atom- and motif-based graphs. In order to boost the representational power of the motif feature, we design a Transformer-based local augmentation module to enrich motif features by introducing heterogeneous atom information in motif representation learning. Besides, we focus on the molecular hierarchical relationship and propose a simple yet effective rescaling module, called contextual self-rescaling, that adaptively recalibrates molecular representations by explicitly modelling interdependencies between atom and motif features. Extensive computational experiments demonstrate that HimGNN can achieve promising performances over state-of-the-art baselines on both classification and regression tasks in molecular property prediction.

Identification and Validation of Ferroptosis-Related Genes in Patients with Acute Spinal Cord Injury.

Ferroptosis plays crucial roles in the pathology of spinal cord injury (SCI). The purpose of this study was to identify differentially expressed ferroptosis-related genes (DE-FRGs) in human acute SCI by bioinformatics analysis and validate the hub DE-FRGs in non-SCI and SCI patients. The GSE151371 dataset was downloaded from the Gene Expression Omnibus and difference analysis was performed. The differentially expressed genes (DEGs) in GSE151371 overlapped with the ferroptosis-related genes (FRGs) obtained from the Ferroptosis Database. A total of 41 DE-FRGs were detected in 38 SCI samples and 10 healthy samples in GSE151371. Then, enrichment analyses of these DE-FRGs were performed for functional annotation. The GO enrichment results showed that upregulated DE-FRGs were mainly associated with reactive oxygen species and redox reactions, and the KEGG enrichment analysis indicated involvement in some diseases and ferroptosis pathways. Protein-protein interaction (PPI) analysis and lncRNA-miRNA-mRNA regulatory network were performed to explore the correlations between genes and regulatory mechanisms. The relationship between DE-FRGs and differentially expressed mitochondria-related genes (DE-MRGs) was also analyzed. Finally, quantitative real-time polymerase chain reaction (qRT-PCR) was used to verify the hub DE-FRGs in clinical blood samples from acute SCI patients and healthy controls. Consistent with the bioinformatics results, qRT-PCR of the clinical samples indicated similar expression levels of TLR4, STAT3, and HMOX1. This study identified DE-FRGs in blood samples from SCI patients, and the results could improve our understanding of the molecular mechanisms of ferroptosis in SCI. These candidate genes and pathways could be therapeutic targets for SCI.

DRLM: A Robust Drug Representation Learning Method and Its Applications.

Learning representations from data is a fundamental step for machine learning. High-quality and robust drug representations can broaden the understanding of pharmacology, and improve the modeling of multiple drug-related prediction tasks, which further facilitates drug development. Although there are a number of models developed for drug representation learning from various data sources, few researches extract drug representations from gene expression profiles. Since gene expression profiles of drug-treated cells are widely used in clinical diagnosis and therapy, it is believed that leveraging them to eliminate cell specificity can promote drug representation learning. In this paper, we propose a three-stage deep learning method for drug representation learning, named DRLM, which integrates gene expression profiles of drug-related cells and the therapeutic use information of drugs. Firstly, we construct a stacked autoencoder to learn low-dimensional compact drug representations. Secondly, we utilize an iterative clustering module to reduce the negative effects of cell specificity and noise in gene expression profiles on the low-dimensional drug representations. Thirdly, a therapeutic use discriminator is designed to incorporate therapeutic use information into the drug representations. The visualization analysis of drug representations demonstrates DRLM can reduce cell specificity and integrate therapeutic use information effectively. Extensive experiments on three types of prediction tasks are conducted based on different drug representations, and they show that the drug representations learned by DRLM outperform other representations in terms of most metrics. The ablation analysis also demonstrates DRLM's effectiveness of merging the gene expression profiles with the therapeutic use information. Furthermore, we input the learned representations into the machine learning models for case studies, which indicates its potential to discover new drug-related relationships in various tasks.

Extracellular vesicle therapy for traumatic central nervous system disorders.

Traumatic central nervous system (CNS) disorders have catastrophic effects on patients, and, currently, there is no effective clinical treatment. Cell transplantation is a common treatment for traumatic CNS injury in animals. In recent years, an increasing number of studies have reported that the beneficial effect of transplanted cells for CNS repair is mediated primarily through the extracellular vesicles (EVs) secreted by the cells, in which microRNAs play a major role. Accordingly, numerous studies have evaluated the roles and applications of EVs secreted by different cell types in neurological diseases. Furthermore, due to their unique biological features, EVs are used as disease biomarkers and drug delivery systems for disease prevention and treatment. We discuss current knowledge related to EVs, focusing on the mechanism underlying their effects on traumatic CNS diseases, and summarize existing research on the potential clinical utility of EVs as disease biomarkers and drug delivery systems.

Restorative therapy using microglial depletion and repopulation for central nervous system injuries and diseases.

Microglia are important resident immune cells in the central nervous system (CNS) and play an important role in its development, homeostasis, and disease treatments. Activated microglia perform diverse functions in mouse models of CNS neurodegenerative diseases or deficits. In humans, microglia have been linked to various neurodegenerative diseases. Following brain or spinal cord injury, microglia express pro- and anti-inflammatory phenotypes at different stages of recovery. With the development of pharmacological and genetic tools for microglial depletion, studies have demonstrated that microglial depletion exerts both positive and negative effects in the treatment of CNS diseases. Notably, microglial depletion provides an empty niche that stimulates production of new microglia. Microglial depletion and repopulation can not only treat diseases by eliminating dysfunctional microglia but can also provide an indication of the molecular mechanisms of diseases. Although this approach has shown impressive results, its use is still in its infancy. In this review, we summarize the current pharmacological and genetic tools for microglial depletion and highlight recent advances in microglial repopulation therapy for the treatment and functional recovery of neurological diseases and deficits. Finally, we briefly discuss the therapeutic challenges and prospective uses of microglial repopulation therapy.

Repair of the Injured Spinal Cord by Schwann Cell Transplantation.

Spinal cord injury (SCI) can result in sensorimotor impairments or disability. Studies of the cellular response to SCI have increased our understanding of nerve regenerative failure following spinal cord trauma. Biological, engineering and rehabilitation strategies for repairing the injured spinal cord have shown impressive results in SCI models of both rodents and non-human primates. Cell transplantation, in particular, is becoming a highly promising approach due to the cells' capacity to provide multiple benefits at the molecular, cellular, and circuit levels. While various cell types have been investigated, we focus on the use of Schwann cells (SCs) to promote SCI repair in this review. Transplantation of SCs promotes functional recovery in animal models and is safe for use in humans with subacute SCI. The rationales for the therapeutic use of SCs for SCI include enhancement of axon regeneration, remyelination of newborn or sparing axons, regulation of the inflammatory response, and maintenance of the survival of damaged tissue. However, little is known about the molecular mechanisms by which transplanted SCs exert a reparative effect on SCI. Moreover, SC-based therapeutic strategies face considerable challenges in preclinical studies. These issues must be clarified to make SC transplantation a feasible clinical option. In this review, we summarize the recent advances in SC transplantation for SCI, and highlight proposed mechanisms and challenges of SC-mediated therapy. The sparse information available on SC clinical application in patients with SCI is also discussed.

Relationship between proprioception and balance control among Chinese senior older adults.

Background: Balance impairment is the most common risk factor for falls among older adults, with three potential factors (tactile sensation, proprioception, and muscle strength) being responsible for their balance control. However, controversies remain on whether or not balance control is related to the three contributors among older adults. Therefore, clarifying the above questions helps explain the mechanisms of increased falls among senior older adults. This study compares the balance control and the three factors and investigates their relationships among older adults of different ages. Methods: 166 participants ultimately passed the qualification assessment and were categorized into younger (YG, 60-69 years, n = 56), middle (MG, 70-79 years, n = 57), or older (OG, ≥80 years, n = 53) aged groups. Berg Balance Scale (BBS) performance, tactile sensation, proprioception, and muscle strength were tested. One-way ANOVA and partial correlation were performed to explore the differences between groups in BBS and its three potential contributors, along with the correlations between them within each age group. Results: Significant differences among the three groups were detected in BBS scores (p < 0.001), tactile sensation at the great toe (p = 0.015) and heel (p = 0.025), proprioception of knee flexion (p < 0.001) and extension (p < 0.001), and ankle plantarflexion (p < 0.001) and dorsiflexion (p < 0.001), and muscle strength of ankle plantarflexion (p < 0.001) and dorsiflexion (p < 0.001), and hip abduction (p < 0.001). Proprioception of knee flexion (r = -0.351, p = 0.009) and extension (r = -0.276, p = 0.041), and ankle plantarflexion (r = -0.283, p = 0.036), and muscle strength of ankle plantarflexion (r = 0.326, p = 0.015) and hip abduction (r = 0.415, p = 0.002) were correlated with BBS among the YG. Proprioception of ankle plantarflexion (r = -0.291, p = 0.030) and muscle strength of ankle plantarflexion (r = 0.448, p = 0.001) and dorsiflexion (r = 0.356, p = 0.007) were correlated with BBS among the MG. Muscle strength of ankle plantarflexion (r = 0.276, p = 0.039) and hip abduction (r = 0.324, p = 0.015) were correlated with BBS among the OG. Conclusion: YG and MG had better balance control, tactile sensation, proprioception, and muscle strength compared to OG. Proprioception correlated with balance control in YG and MG, but not in the OG. The worsen proprioception among the OG could be the key for increased falls. Exercise should be recommended to improve proprioception among senior older adults.

Effects of traditional Chinese herb hot compress combined with therapeutic exercise on pain, proprioception, and functional performance among older adults with knee osteoarthritis: A randomized controlled trial.

Background: Knee osteoarthritis (KOA) is one of the most common chronic progressive diseases with degenerative destruction of articular cartilage and bone, leading to knee pain, impaired proprioception, and reduced functional performance. This study was to investigate the effects of an 8-week Traditional Chinese herb hot compress (TCHHC) combined with therapeutic exercise (TE) on pain, proprioception, and functional performance among older adults with KOA. Methods: Twenty-seven older adults with KOA were recruited and randomly assigned to the TCHHC + TE or TE groups. Thirteen participants received TCHHC + TE, and fourteen received TE. At pre- (week 0) and post-intervention (week 9), their pain, joint proprioception, and functional performance were measured. Two-way ANOVA with repeated measures was adopted to analyze the data. Results: Compared with week 0, the pain score, proprioception thresholds of knee extension and ankle plantarflexion, and the times of TUG and 20-m walk tests decreased more significantly in the TCHHC + TE group than in the TE group at week 9. Conclusion: Compared with TE, the 8-week TCHHC + TE was superior in relieving pain, recovering proprioception, and improving functional performance among older adults with KOA. It is recommended that TCHHC should be adopted prior to TE to enhance the effects of KOA rehabilitation.

GraphCDR: a graph neural network method with contrastive learning for cancer drug response prediction.

Predicting the response of a cancer cell line to a therapeutic drug is an important topic in modern oncology that can help personalized treatment for cancers. Although numerous machine learning methods have been developed for cancer drug response (CDR) prediction, integrating diverse information about cancer cell lines, drugs and their known responses still remains a great challenge. In this paper, we propose a graph neural network method with contrastive learning for CDR prediction. GraphCDR constructs a graph neural network based on multi-omics profiles of cancer cell lines, the chemical structure of drugs and known cancer cell line-drug responses for CDR prediction, while a contrastive learning task is presented as a regularizer within a multi-task learning paradigm to enhance the generalization ability. In the computational experiments, GraphCDR outperforms state-of-the-art methods under different experimental configurations, and the ablation study reveals the key components of GraphCDR: biological features, known cancer cell line-drug responses and contrastive learning are important for the high-accuracy CDR prediction. The experimental analyses imply the predictive power of GraphCDR and its potential value in guiding anti-cancer drug selection.

MVGCN: data integration through multi-view graph convolutional network for predicting links in biomedical bipartite networks.

MOTIVATION: There are various interaction/association bipartite networks in biomolecular systems. Identifying unobserved links in biomedical bipartite networks helps to understand the underlying molecular mechanisms of human complex diseases and thus benefits the diagnosis and treatment of diseases. Although a great number of computational methods have been proposed to predict links in biomedical bipartite networks, most of them heavily depend on features and structures involving the bioentities in one specific bipartite network, which limits the generalization capacity of applying the models to other bipartite networks. Meanwhile, bioentities usually have multiple features, and how to leverage them has also been challenging. RESULTS: In this study, we propose a novel multi-view graph convolution network (MVGCN) framework for link prediction in biomedical bipartite networks. We first construct a multi-view heterogeneous network (MVHN) by combining the similarity networks with the biomedical bipartite network, and then perform a self-supervised learning strategy on the bipartite network to obtain node attributes as initial embeddings. Further, a neighborhood information aggregation (NIA) layer is designed for iteratively updating the embeddings of nodes by aggregating information from inter- and intra-domain neighbors in every view of the MVHN. Next, we combine embeddings of multiple NIA layers in each view, and integrate multiple views to obtain the final node embeddings, which are then fed into a discriminator to predict the existence of links. Extensive experiments show MVGCN performs better than or on par with baseline methods and has the generalization capacity on six benchmark datasets involving three typical tasks. AVAILABILITY AND IMPLEMENTATION: Source code and data can be downloaded from https://github.com/fuhaitao95/MVGCN. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.