Cancer drug response prediction with surrogate modeling-based graph neural architecture search.

MOTIVATION: Understanding drug-response differences in cancer treatments is one of the most challenging aspects of personalized medicine. Recently, graph neural networks (GNNs) have become state-of-the-art methods in many graph representation learning scenarios in bioinformatics. However, building an optimal handcrafted GNN model for a particular drug sensitivity dataset requires manual design and fine-tuning of the hyperparameters for the GNN model, which is time-consuming and requires expert knowledge. RESULTS: In this work, we propose AutoCDRP, a novel framework for automated cancer drug-response predictor using GNNs. Our approach leverages surrogate modeling to efficiently search for the most effective GNN architecture. AutoCDRP uses a surrogate model to predict the performance of GNN architectures sampled from a search space, allowing it to select the optimal architecture based on evaluation performance. Hence, AutoCDRP can efficiently identify the optimal GNN architecture by exploring the performance of all GNN architectures in the search space. Through comprehensive experiments on two benchmark datasets, we demonstrate that the GNN architecture generated by AutoCDRP surpasses state-of-the-art designs. Notably, the optimal GNN architecture identified by AutoCDRP consistently outperforms the best baseline architecture from the first epoch, providing further evidence of its effectiveness. AVAILABILITY AND IMPLEMENTATION: https://github.com/BeObm/AutoCDRP.

MTGNN: Multi-Task Graph Neural Network based few-shot learning for disease similarity measurement.

Similar diseases are usually caused by molecular origins or similar phenotypes. Confirming the relationship between diseases can help researchers gain a deep insight of the pathogenic mechanisms of emerging complex diseases, and improve the corresponding diagnoses and treatment. Therefore, similar diseases are considerably important in biology and pathology. However, the insufficient number of labelled similar disease pairs cannot support the optimal training of the models. In this paper, we propose a Multi-Task Graph Neural Network (MTGNN) framework to measure disease similarity by few-shot learning. To tackle the problem of insufficient number of labelled similar disease pairs, we design the multi-task optimization strategy to train the graph neural network for disease similarity task (lack of labelled training data) by introducing link prediction task (sufficient labelled training data). The similarity between diseases can then be obtained by measuring the distance between disease embeddings in high-dimensional space learning from the double tasks. The experiment results evaluate the performance of MTGNN and illustrate its advantages over previous methods on few labeled training dataset.

The developing Human Connectome Project (dHCP) automated resting-state functional processing framework for newborn infants.

The developing Human Connectome Project (dHCP) aims to create a detailed 4-dimensional connectome of early life spanning 20-45 weeks post-menstrual age. This is being achieved through the acquisition of multi-modal MRI data from over 1000 in- and ex-utero subjects combined with the development of optimised pre-processing pipelines. In this paper we present an automated and robust pipeline to minimally pre-process highly confounded neonatal resting-state fMRI data, robustly, with low failure rates and high quality-assurance. The pipeline has been designed to specifically address the challenges that neonatal data presents including low and variable contrast and high levels of head motion. We provide a detailed description and evaluation of the pipeline which includes integrated slice-to-volume motion correction and dynamic susceptibility distortion correction, a robust multimodal registration approach, bespoke ICA-based denoising, and an automated QC framework. We assess these components on a large cohort of dHCP subjects and demonstrate that processing refinements integrated into the pipeline provide substantial reduction in movement related distortions, resulting in significant improvements in SNR, and detection of high quality RSNs from neonates.

ConnectedAlign: a PPI network alignment method for identifying conserved protein complexes across multiple species.

BACKGROUND: In bioinformatics, network alignment algorithms have been applied to protein-protein interaction (PPI) networks to discover evolutionary conserved substructures at the system level. However, most previous methods aim to maximize the similarity of aligned proteins in pairwise networks, while concerning little about the feature of connectivity in these substructures, such as the protein complexes. RESULTS: In this paper, we identify the problem of finding conserved protein complexes, which requires the aligned proteins in a PPI network to form a connected subnetwork. By taking the feature of connectivity into consideration, we propose ConnectedAlign, an efficient method to find conserved protein complexes from multiple PPI networks. The proposed method improves the coverage significantly without compromising of the consistency in the aligned results. In this way, the knowledge of protein complexes in well-studied species can be extended to that of poor-studied species. CONCLUSIONS: We conducted extensive experiments on real PPI networks of four species, including human, yeast, fruit fly and worm. The experimental results demonstrate dominant benefits of the proposed method in finding protein complexes across multiple species.

RCSS: A Real-Time On-Demand Charging Scheduling Scheme for Wireless Rechargeable Sensor Networks.

With the emergence of edge computing, a large number of devices such as sensor nodes have been deployed in the edge network to sense and process data. However, how to provide real-time on-demand energy for these edge devices is a new challenge issue of edge networks. In real-world applications of edge computing, sensor nodes usually have different task burdens due to the environmental impact, which results in a dynamic change of the energy consumption rate at different nodes. Therefore, the traditional periodical charging mode cannot meet the nodes charging demand that have dynamic energy consumption. In this paper, we propose a real-time on-demand charging scheduling scheme (RCSS) under the condition of limited mobile charger capacity. In the process of building the charging path, RCSS adequately considers the dynamic energy consumption of different node, and puts forward the next node selection algorithm. At the same time, a method to determine the feasibility of charging circuit is also proposed to ensure the charging efficiency. During the charging process, RCSS is based on adaptive charging threshold to reduce node mortality. Compared with existing approaches, the proposed RCSS achieves better performance in the number of survival nodes, the average service time and charging efficiency.

AutoAMS: Automated attention-based multi-modal graph learning architecture search.

Multi-modal attention mechanisms have been successfully used in multi-modal graph learning for various tasks. However, existing attention-based multi-modal graph learning (AMGL) architectures heavily rely on manual design, requiring huge effort and expert experience. Meanwhile, graph neural architecture search (GNAS) has made great progress toward automatically designing graph-based learning architectures. However, it is challenging to directly adopt existing GNAS methods to search for better AMGL architectures because of the search spaces that only focus on designing graph neural network architectures and the search objective that ignores multi-modal interactive information between modalities and long-term content dependencies within different modalities. To address these issues, we propose an automated attention-based multi-modal graph learning architecture search (AutoAMS) framework, which can automatically design the optimal AMGL architectures for different multi-modal tasks. Specifically, we design an effective attention-based multi-modal (AM) search space consisting of four sub-spaces, which can jointly support the automatic search of multi-modal attention representation and other components of multi-modal graph learning architecture. In addition, a novel search objective based on an unsupervised multi-modal reconstruction loss and task-specific loss is introduced to search and train AMGL architectures. The search objective can extract the global features and capture multi-modal interactions from multiple modalities. The experimental results on multi-modal tasks show strong evidence that AutoAMS is capable of designing high-performance AMGL architectures.

AutoDDI: Drug-Drug Interaction Prediction With Automated Graph Neural Network.

Drug-drug interaction (DDI) has attracted widespread attention because when incompatible drugs are taken together, DDI will lead to adverse effects on the body, such as drug poisoning or reduced drug efficacy. The adverse effects of DDI are closely determined by the molecular structures of the drugs involved. To represent drug data effectively, researchers usually treat the molecular structure of drugs as a molecule graph. Then, previous studies can use the handcrafted graph neural network (GNN) model to learn the molecular graph representations of drugs for DDI prediction. However, in the field of bioinformatics, manually designing GNN architectures for specific molecular structure datasets is time-consuming and depends on expert experience. To address this problem, we propose an automatic drug-drug interaction prediction method named AutoDDI that can efficiently and automatically design the GNN architecture for drug-drug interaction prediction without manual intervention. To this end, we first design an effective search space for drug-drug interaction prediction by revisiting various handcrafted GNN architectures. Then, to efficiently and automatically design the optimal GNN architecture for each drug dataset from the search space, a reinforcement learning search algorithm is adopted. The experiment results show that AutoDDI can achieve the best performance on two real-world datasets. Moreover, the visual interpretation results of the case study show that AutoDDI can effectively capture drug substructure for drug-drug interaction prediction.

Transparent multifunctional cellulose-based conductive hydrogel for wearable strain sensors and arrays.

Conductive hydrogels as promising candidate materials for flexible strain sensors have gained considerable attentions. However, it is still a great challenge to construct hydrogel with multifunctional performance via natural polymer. Herein, a novel multifunctional conductive hydrogel based on methylcellulose and cellulose nanocrystal was prepared via a facile and low-cost strategy. Methylcellulose (MC) was introduced to not only guarantee the stability of tannic acid coated cellulose nanocrystal (TA@CNCs) in LiCl solution, but also improve anti-freezing ability. The obtained hydrogel exhibited high transparency (98 % at 800 nm), good stretchability (663.1 %), low temperature tolerance (-23.9 °C), superior conductivity (2.89 S/m) and excellent UV shielding behavior. Flexible strain sensor assembled by the prepared hydrogels can be used to detect human body motions include subtle and large motions, and exhibited good sensitivity and stability over a wide temperature range. Multiple flexible hydrogels can also be assembled into a 3D sensor array to detect the distribution and magnitude of spatial pressure. Therefore, the hydrogels prepared via natural polymers will have broad application prospects in wearable devices, electronic skin and multifunctional sensor components.

Teleost Muc2 and Muc5ac: Key guardians of mucosal immunity in flounder (Paralichthys olivaceus).

Mucins secreted by mucous cells constitute a core part of the defense line against the invasion of pathogens. However, mucins' structure and immunological functions remain largely unknown in teleost. In this study, two typical mucins, Muc2 and Muc5ac of flounder (Paralichthys olivaceus), were cloned and their physicochemical properties, structure and conservation were analyzed. Notably, specific antibodies against flounder Muc2 and Muc5ac were developed. It was verified at the gene and protein level that Muc2 was expressed in the hindgut and gills but not in the skin, while Muc5ac was expressed in the skin and gills but not in the hindgut. After flounders were immunized by immersion with inactivated Edwardsiella tarda, Muc2 and Muc5ac were significantly upregulated at both the gene expression and protein secretion levels, and Muc2+/Muc5ac+ mucous cells proliferated and increased secretion of Muc2 and Muc5ac. Moreover, Muc2 and Muc5ac exerted retention and clearance effects on E. tarda in a short period (within 1 dpi). These results revealed the characterization of fish mucins Muc2 and Muc5ac at the protein level and clarified the role of mucins as key guardians to maintain the mucus barrier, which advanced our understanding of teleost mucosal barrier.

Sclerostin Transduced Bone Marrow Mesenchymal Stem Cells Promote Fracture Healing in Rats Through the Wnt/ß-Catenin Signal Pathway.

The prognosis of fracture is directly related to several factors. Due to the limitations of existing treatment strategies, there are still many fractures with poor healing. Bone marrow mesenchymal stem cells (BMSCs) have the potential to differentiate into osteoblasts and chondrocytes. Therefore, BMSC transplantation is promised as an effective method for treating bone fractures. We aim to explore whether silently expressing sclerostin gene (SOST) can promote bone formation through the SOST/Wnt/ß-catenin signal pathway. We isolated rat BMSCs and the target gene (SOST shRNA) was transduced into them for osteogenic induction. The results showed that SOST significantly inhibited the proliferation and osteogenic differentiation of BMSCs during osteogenic induction, whereas silently expressing SOST not only increased the number of surviving BMSCs but also promoted the expression of osteogenesis-related proteins RUNX2, osteoprotegerin, Collagen I (COL-I), and bone morphogenetic protein-2 during osteogenic induction. The results of imaging examination in rats show that downregulating the expression of SOST can promote the formation of bony callus and the transformation of cartilage tissue into normal bone tissue, and then accelerate the healing of osteoporotic fracture. In addition, we also found that SOST silencing can activate the Wnt/ß-catenin pathway to achieve these effects. In conclusion, SOST silencing can promote the proliferation and osteogenic differentiation of BMSCs in situ, and therefore may enhance the therapeutic efficiency of BMSC transplantation in OPF.

Dysregulation of interleukin-8 is involved in the onset and relapse of schizophrenia: An independent validation and meta-analysis.

As a major mental health disorder, symptoms of schizophrenia (SCZ) include delusions, reduced motivation, hallucinations, reduced motivation and a variety of cognitive disabilities. Many of these symptoms are now known to be associated with abnormal regulation of the immune system. Low blood levels of cytokines and chemokines have been suggested to be one of the underlying causes of SCZ. However, their biological roles at different stages of SCZ remain unclear. Our objective was to investigate expression patterns of cytokines and chemokines at different stages of onset and relapse in SCZ patients and to conduct an analysis of their relationship to disease progression. We also aimed to identify immune features associated with different disease trajectories in patients with SCZ. Gene set enrichment analysis (GSEA) was used to interrogate the GSE27383 dataset and identify key genes associated with inflammation. These results led us to recruit 36 healthy controls, 40 patients with first-episode psychosis (FEP), and 39 patients with SCZ relapse. Meso Scale Discovery technology was used to independently validate serum levels of 35 cytokines and chemokines. This was followed by a meta-analysis to gain a more comprehensive understanding of the role of interleukin-8 (IL-8/CXCL8) in SCZ. Analysis of the GSE27383 database revealed 3596 genes with distinct expression patterns. A significant portion of these genes were identified as inflammation-related and showed remarkable enrichment in three key pathways: IL-17, cytokine-cytokine receptor, and AGE-RAGE signaling in diabetic complications. We observed co-expression of CXCL8 and IL-16 within these three pathways. In a subsequent analysis of independently validated samples, a notable discrepancy was detected in the inflammatory status between individuals experiencing FEP and those in relapse. In particular, expression of CXCL8 demonstrated superior predictive capability in FEP and relapsed patients. Notably, results of the meta-analysis confirmed that Chinese and European populations were consistent with the overall results (Z = 4.60, P < 0.001; Z = 3.70, P < 0.001). However, in the American subgroup, there was no significant difference in CXCL8 levels between patients with SCZ compared to healthy controls (Z = 1.09, P = 0.277). Our findings suggest that the inflammatory response in patients with SCZ differs across the different stages, with CXCL8 emerging as a potential predictive factor. Collectively, our data suggest that CXCL8 has the potential to serve as a significant immunological signature of SCZ subtypes. Trial registration: The clinical registration number for this trial is ChiCTR2100045240 (Registration Date: 2021/04/09).

Multi-View Graph Neural Architecture Search for Biomedical Entity and Relation Extraction.

Recently, graph neural architecture search (GNAS) frameworks have been successfully used to automatically design the optimal neural architectures for many problems such as node classification and graph classification. In the existing GNAS frameworks, the designed graph neural network (GNN) architectures learn the representation of homogenous graphs with one type of relationship connecting two nodes. However, multi-view graphs, where each view represents a type of relationship among nodes, are ubiquitous in the real world. The traditional GNAS frameworks learn the graph representation without considering the interactions between nodes and multiple relationships, so they fail to solve multi-view graph-based problems, such as multi-view graphs modelling the biomedical entity and relation extraction tasks. In this paper, we propose MVGNAS, a multi-view graph neural network automatic modelling framework for biomedical entity and relation extraction, to resolve this challenge. In MVGNAS, we propose an automatic multi-view representation learning to learn low-dimensional representations of nodes that capture multiple relationships in a multi-view graph, representing the first research work in literature to solve the problem of multi-view graph representation learning architecture search for biomedical entity and relation extraction tasks. The experimental results demonstrate that MVGNAS can achieve the best performance in biomedical entity and relation extraction tasks against the state-of-the-art baseline methods.

Surface modification of chitin nanofibers with dopamine as efficient nanosorbents for enhanced removal of dye pollution and metal ions.

The development of environmentally friendly and low-cost adsorbents with high adsorption capacity remains a challenge. Herein, chitin nanofiber-polydopamine composite materials (CNDA) have been obtained by surface modification of chitin nanofiber using dopamine. According to the results of transmission electron microscopy (TEM), X-ray diffraction (XRD), Fourier Transform Infrared Spectrometer (FTIR), and X-ray photoelectron spectrometer (XPS), polydopamine have been successfully coated on the surface of chitin nanofiber (ChNF). The ability to remove methylene blue (MB) has been analyzed via standard adsorption experiments, indicating that the maximum adsorption capacity (qmax) can reach 196.6 mg/g at MB initial concentration of 50 mg/L. Most importantly, the adsorption kinetics, isotherm, and thermodynamics were used to investigate the MB adsorption mechanism on composites. This indicated that the polydopamine on the surface of chitin nanofiber (ChNF) plays an important role in the MB dye adsorption. Moreover, the removal ability of CNDA to metal ions has also been investigated, indicating high capacities for Fe3+, Mn2+, Cu2+, and Ni2+. Based on their biodegradability and good adsorption capacity, the CNDA composite material can be considered a promising adsorbent for wastewater treatment.

[Corrigendum] Icaritin promotes the osteogenesis of bone marrow mesenchymal stem cells via the regulation of sclerostin expression.

Following the publication of the above article, the authors have contacted the Editorial Office to explain that they had assembled the cellular morphological images in Fig. 1A on p. 819 incorrectly; essentially, the cell morphology of 2 passages of hBMSCs (centre panel) should have been shown as the data panel for 3 passages of hBMSCs (right-hand panel), and likewise, the cell morphology of 3 passages of hBMSCs should have been shown as the data panel for 2 passages of hBMSCs. The revised version of Fig. 1 is shown below. The authors confirm that the errors associated with this figure did not have any significant impact on either the results or the conclusions reported in this study, and are grateful to the Editor of International Journal of Molecular Medicine for allowing them the opportunity to publish this Corrigendum. Furthermore, they apologize to the readership of the Journal for any inconvenience caused. [International Journal of Molecular Medicine 45: 816-824, 2020; DOI: 10.3892/ijmm.2020.4470].

Intermediary roles of prospective memory and retrospective memory in the comorbidity of depression and pain.

Background: Patients who suffer comorbidity of major depressive disorder (MDD) and chronic pain (CP) maintain a complex interplay between maladaptive prospective memory (PM) and retrospective memory (RM) with physical pain, and their complications are still unknown. Aims: We aimed to focus on the full cognitive performance and memory complaints in patients with MDD and CP, patients with depression without CP, and control subjects, considering the possible influence of depressed affect and chronic pain severity. Methods: According to the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders and the criteria given by the International Association of Pain, a total of 124 participants were included in this cross-sectional cohort study. Among them, 82 depressed inpatients and outpatients from Anhui Mental Health centre were divided into two groups: a comorbidity group(patients with MDD and CP) (n=40) and a depression group (patients with depression without CP) (n=42). Meanwhile, 42 healthy controls were screened from the hospital's physical examination centre from January 2019 to January 2022. The Hamilton Depression Rating Scale-24 (HAMD-24) and Beck Depression Inventory-II (BDI-II) were used to evaluate the severity of depression. The Pain Intensity Numerical Rating Scale (PI-NRS), Short-Form McGill Pain Questionnaire-2 Chinese version (SF-MPQ-2-CN), Montreal Cognitive Assessment-Basic Section (MoCA-BC), and Prospective and Retrospective Memory Questionnaire (PRMQ) were used to assess pain-related features and the global cognitive functioning of study participants. Results: The impairments in PM and RM differed remarkably among the three groups (F=7.221, p<0.001; F=7.408, p<0.001) and were severe in the comorbidity group. Spearman correlation analysis revealed the PM and RM were positively correlated with continuous pain and neuropathic pain (r=0.431, p<0.001; r=0.253, p=0.022 and r=0.415, p<0.001; r=0.247, p=0.025), respectively. Regression analysis indicated a significant positive relationship between affective descriptors and total BDI-II score (ß=0.594, t=6.600, p<0.001). Examining the mediator pathways revealed the indirect role of PM and RM in patients with comorbid MDD and CP. Conclusions: Patients with comorbid MDD and CP presented more PM and RM impairments than patients with MDD without CP. PM and RM are possibly mediating factors that affect the aetiology of comorbid MDD and CP. Trial registration number: chiCTR2000029917.

Automatic Thoughts, Self-Stigma, and Resilience Among Schizophrenia Patients with Metabolic Syndrome: A Cross-Sectional Study.

Purpose: The study aims to clarify the negative psychological state and resilience impairments of schizophrenia (SCZ) with metabolic syndrome (MetS) while evaluating their potential as risk factors. Patients and Methods: We recruited 143 individuals and divided them into three groups. Participants were evaluated using the Positive and Negative Syndrome Scale (PANSS), Hamilton Depression Rating Scale (HAMD)-24, Hamilton Anxiety Rating Scale (HAMA)-14, Automatic Thoughts Questionnaire (ATQ), Stigma of Mental Illness scale and Connor-Davidson Resilience Scale (CD-RISC). Serum biochemical parameters were measured by automatic biochemistry analyzer. Results: The score of ATQ was highest in the MetS group (F = 14.5, p < 0.001), and the total score of CD-RISC, subscale tenacity score and subscale strength score of CD-RISC were lowest in the MetS group (F = 8.54, p < 0.001; F = 5.79, p = 0.004; F = 10.9, p < 0.001). A stepwise regression analysis demonstrated that a negative correlation was observed among the ATQ with employment status, high-density lipoprotein (HDL-C), and CD-RISC (ß=-0.190, t=-2.297, p = 0.023; ß=-0.278, t=-3.437, p = 0.001; ß=-0.238, t=-2.904, p = 0.004). A positive correlation was observed among the ATQ with waist, TG, WBC, and stigma (ß=0.271, t = 3.340, p = 0.001; ß=0.283, t = 3.509, p = 0.001; ß=0.231, t = 2.815, p = 0.006; ß=0.251, t=-2.504, p = 0.014). The area under the receiver-operating characteristic curve analysis showed that among all independent predictors of ATQ, the TG, waist, HDL-C, CD-RISC, and stigma presented excellent specificity at 0.918, 0.852, 0.759, 0.633, and 0.605, respectively. Conclusion: Results suggested that the non-MetS and MetS groups had grievous sense of stigma, particularly, high degree of ATQ and resilience impairment was shown by the MetS group. The TG, waist, HDL-C of metabolic parameters, CD-RISC, and stigma presented excellent specificity to predict ATQ, and the waist showed excellent specificity to predict low resilience level.

Effect of Mixed Acid Fluid on the Pore Structure of High Rank Coal and Acid Fluid Optimization.

Acidizing technology is an important means to increase production in oil-gas reservoirs. In recent years, acidizing technology has been widely used to increase the permeability of coal seams to enhance gas extraction, where acidizing fluid is the key factor to determine the permeability improvement effect by acidizing technology. In order to clarify the influence of mixed acid fluid on the pore structure of high rank coal and seek the optimal mixed acid fluid suitable for acidizing and permeability improvement of high rank coal in the Jiaozuo coal mine area. Taking the Jiulishan Mine in the Jiaozuo mining area as an example, low field nuclear magnetic resonance (LFNMR) test and static dissolution test were conducted to obtain the T 2 spectrum, porosity, movable fluid saturation, pore throat distribution, nuclear magnetic permeability, and dissolution rate of coal samples before and after treatment with distilled water and three mixed acid fluids. On this basis, the influence of mixed acid fluid on the pore structure of high rank coal was analyzed and the optimal mixed acid fluid suitable for high rank coal was selected. The results showed that the pore size, number, and volume of all kinds of pore sizes of coal samples treated with distilled water all decreased, which was manifested by the decrease of effective porosity and nuclear magnetic permeability. After acidification, the proportion of micropore volume in coal decreased significantly, the number and proportion of pore volume of mesopores and macropore-microfractures increased significantly, and the connectivity between mesopores and macropore-microfractures was enhanced, which was characterized by the increase in effective porosity and nuclear magnetic permeability of coal samples. After acidification, the pore-throat ratio of adsorption pores of all coal samples decreased, while the pore-throat ratio of seepage pores increased. By comparatively analyzing the change law of pore structure of coal samples before and after acidizing with three kinds of mixed acid fluids, the optimal mixed acid fluid suitable for acidizing and permeability improvement of high rank coal in the Jiaozuo coal mine area was selected, which was 12%HCL +3%HF.

AutoMSR: Auto Molecular Structure Representation Learning for Multi-label Metabolic Pathway Prediction.

It is significant to comprehend the relationship between metabolic pathway and molecular pathway for synthesizing new molecules, for instance optimizing drug metabolization. In bioinformatics fields, multi-label prediction of metabolic pathways is a typical manner to understand this relationship. Graph neural networks (GNNs) have become an effective method to extract molecular structure's features for multi-label prediction of metabolic pathways. Though GNNs can effectively capture structural features from molecular structure graphs, building a well-performed GNN model for a given molecular structure data set requires the manual design of the GNN architecture and fine-tuning of the hyperparameters, which are time-consuming and rely on expert experience. To address the above challenge, we design an end-to-end automatic molecular structure representation learning framework named AutoMSR that can design the optimal GNN model based on a given molecular structure data set without manual intervention. We propose a multi-seed age evolution (MSAE) search algorithm to identify the optimal GNN architecture from the GNN architecture subspace. For a given molecular structure data set, AutoMSR first uses MSAE to search the GNN architecture, and then it adopts a tree-structured parzen estimator to obtain the best hyperparameters in the hyperparameters subspace. Finally, AutoMSR automatically constructs the optimal GNN model based on the best GNN architecture and hyperparameters to extract the molecular structure features for multi-label metabolic pathway prediction. We test the performance of AutoMSR on the real data set KEGG. The experiment results show that AutoMSR outperforms baseline methods on different multi-label classification evaluation metrics.

Predicting the Survival of Cancer Patients With Multimodal Graph Neural Network.

In recent years, cancer patients survival prediction holds important significance for worldwide health problems, and has gained many researchers attention in medical information communities. Cancer patients survival prediction can be seen the classification work which is a meaningful and challenging task. Nevertheless, research in this field is still limited. In this work, we design a novel Multimodal Graph Neural Network (MGNN)framework for predicting cancer survival, which explores the features of real-world multimodal data such as gene expression, copy number alteration and clinical data in a unified framework. Specifically, we first construct the bipartite graphs between patients and multimodal data to explore the inherent relation. Subsequently, the embedding of each patient on different bipartite graphs is obtained with graph neural network. Finally, a multimodal fusion neural layer is proposed to fuse the medical features from different modality data. Comprehensive experiments have been conducted on real-world datasets, which demonstrate the superiority of our modal with significant improvements against state-of-the-arts. Furthermore, the proposed MGNN is validated to be more robust on other four cancer datasets.

Network Association of Biochemical and Inflammatory Abnormalities With Psychiatric Symptoms in First-Episode Schizophrenia Patients.

Background: Cardiovascular disease (CVD) risk factors such as dyslipidemia and systemic aberrant inflammatory processes may occur in patients with psychotic disorders, which may cause increased mortality. The interplay between immune and metabolic markers and its contribution to the clinical symptoms of schizophrenia (SCZ) remain unclear. This study aimed to examine the association of a series of inflammatory factors, plasma biochemical indicators, and SCZ clinical symptomatology with the severity of SCZ symptoms. Methods: A total of 115 participants, including 79 first-episode drug-naïve patients with SCZ and 36 healthy controls, were enrolled in this study. Semi-structured interviews were used to collect sociodemographic data, family history of SCZ, and medical and psychiatric history. The Brief Psychiatric Rating Scale (BPRS) and the Positive and Negative Syndrome Scale (PANSS) were administered by a clinical psychiatrist to evaluate the symptom severity of patients with SCZ. Plasma inflammatory cytokines were measured by a fully automated electrochemiluminescent immunoassay (Meso Scale Discovery). Results: Blood routine, biochemical, and inflammation cytokine test results showed that the levels of white blood cell count, neutrophil count, natrium, CRP, IL-8, IL-6, IL-13, and IL-16 significantly increased in the case group than in the healthy controls (p < 0.05), whereas levels of red blood cell count, hemoglobin concentration, mean corpuscular hemoglobin concentration, total protein, albumin, total bile acid, high-density lipoprotein (HDL), apolipoprotein A1, blood urea nitrogen, kalium and IL-15 were lower than in the healthy controls (p < 0.05). Correlation network analysis results shown that the natrium, HDL and red blood cell count were the top 3 factors closely to with BPRS and PANSS related clinical symptoms among of correlation network (degree = 4). ROC curve analysis explored the IL-16, IL-8, IL-13, IL-15, natrium, and HDL had highly sensitivity and specificity to the predictive validity and effectiveness for SCZ symptoms. Conclusion: Our study revealed a complex interactive network correlation among the cardiovascular risk factors, biological immunity profiles, and psychotic symptoms in first-episode patients. Abnormal inflammatory factors and CVD risk factors had high sensitivity and specificity for predicting SCZ symptoms. Generally, our study provided novel information on the immune-related mechanisms involved in early CVD risk in patients with psychotic disorders.