Search

Identifying native-like protein-ligand complexes (PLCs) from an abundance of docking decoys is critical for large-scale virtual drug screening in early-stage drug discovery lead searching efforts. Providing reliable prediction is still a challenge for most current affinity predicting models because of a lack of non-binding data during model training, lost critical physical-chemical features, and difficulties in learning abstract information with limited neural layers. In this work, we proposed a deep learning model, DeepBindBC, for classifying putative ligands as binding or non-binding. Our model incorporates information on non-binding interactions, making it more suitable for real applications. ResNet model architecture and more detailed atom type representation guarantee implicit features can be learned more accurately. Here, we show that DeepBindBC outperforms Autodock Vina, Pafnucy, and DLSCORE for three DUD.E testing sets. Moreover, DeepBindBC identified a novel human pancreatic α-amylase binder validated by a fluorescence spectral experiment (K

Microfluidics is characterized by the manipulation of fluids in submillimeter channels and has great application potential in encapsulation. To further extend the application of microfluidics in food industries, a 3D printed microfluidic device is used to encapsulate vitamin A and improve its stability. Two natural macromolecules, sodium alginate and gelatin, are added to water as the continuous phase to generate monodisperse emulsion. Under different flow rate ratios, the diameter of droplets decreases with the increase of continuous flow rate. However, at the same flow rate ratio, varying the dispersed and continuous flow rates does not significantly change the diameter and size distribution of emulsion collected. The prepared O/W (oil/water) single emulsion can form microgel particles and avoid degradation of vitamin A by simulated gastric acid; the encapsulated vitamin A will not be released until particles reach simulated intestinal tract. In the simulated digestion in vitro, no vitamin A is released for 2 h in the acidic environment; under an alkaline or neutral environment such as those in intestinal fluids, vitamin A can be released from the microgel particles within 2.5 h. Using the presented approach, emulsions encapsulating vitamin A have been prepared and can potentially be applied to encapsulate other oil-soluble substances in the food industry.

Model compression is a promising approach for reducing the model size of pretrained-language-models (PLMs) on low resource edge devices and applications. Unfortunately, the compression process always accompanies a cost of performance degradation, especially for the low resource downstream tasks, i.e., multiple-choice question answering. To address the degradation issue of model compression on PLMs, we proposed an end-to-end reptile (ETER) meta-learning approach to improving the performance of PLMs on the low resource multiple-choice question answering task. Specifically, our ETER improves the traditional two-stage meta-learning to an end-to-end manner, integrating the target finetuning stage into the meta training stage. To strengthen the generic meta-learning, ETER employs two-level meta-task construction from instance-level and domain-level to enrich its task generalization. What is more, ETER optimizes meta-learning by parameter constraints to reduce its parameter learning space. Experiments demonstrate that ETER significantly improved the performance of compressed PLMs and achieved large superiority over the baselines on different datasets.

A substantial proportion of patients with unresectable stage III non-small-cell lung cancer (NSCLC) cannot either tolerate or access concurrent chemoradiotherapy, so sequential chemoradiotherapy is commonly used. We assessed the efficacy and safety of sugemalimab, an anti-PD-L1 antibody, in patients with stage III NSCLC whose disease had not progressed after concurrent or sequential chemoradiotherapy.GEMSTONE-301 is a randomised, double-blind, placebo-controlled, phase 3 trial in patients with locally advanced, unresectable, stage III NSCLC, done at 50 hospitals or academic research centres in China. Eligible patients were aged 18 years or older with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 who had not progressed after concurrent or sequential chemoradiotherapy. We randomly assigned patients (2:1, using an interactive voice-web response system) to receive sugemalimab 1200 mg or matching placebo, intravenously every 3 weeks for up to 24 months. Stratification factors were ECOG performance status, previous chemoradiotherapy, and total radiotherapy dose. The investigators, trial coordination staff, patients, and study sponsor were masked to treatment allocation. The primary endpoint was progression-free survival as assessed by blinded independent central review (BICR) in the intention-to-treat population. Safety was assessed in all participants who received at least one dose of assigned study treatment. The study has completed enrolment and the results of a preplanned analysis of the primary endpoint are reported here. The trial is registered with ClinicalTrials.gov, NCT03728556.Between Aug 30, 2018 and Dec 30, 2020, we screened 564 patients of whom 381 were eligible. Study treatment was received by all patients randomly assigned to sugemalimab (n=255) and to placebo (n=126). At data cutoff (March 8, 2021), median follow-up was 14·3 months (IQR 6·4-19·4) for patients in the sugemalimab group and 13·7 months (7·1-18·4) for patients in the placebo group. Progression-free survival assessed by BICR was significantly longer with sugemalimab than with placebo (median 9·0 months [95% CI 8·1-14·1] vs 5·8 months [95% CI 4·2-6·6]; stratified hazard ratio 0·64 [95% CI 0·48-0·85], p=0·0026). Grade 3 or 4 treatment-related adverse events occurred in 22 (9%) of 255 patients in the sugemalimab group versus seven (6%) of 126 patients in the placebo group, the most common being pneumonitis or immune-mediated pneumonitis (seven [3%] of 255 patients in the sugemalimab group vs one [1%] of 126 in the placebo group). Treatment-related serious adverse events occurred in 38 (15%) patients in the sugemalimab group and 12 (10%) in the placebo group. Treatment-related deaths were reported in four (2%) of 255 patients (pneumonia in two patients, pneumonia with immune-mediated pneumonitis in one patient, and acute hepatic failure in one patient) in the sugemalimab group and none in the placebo group.Sugemalimab after definitive concurrent or sequential chemoradiotherapy could be an effective consolidation therapy for patients with stage III NSCLC whose disease has not progressed after sequential or concurrent chemoradiotherapy. Longer follow-up is needed to confirm this conclusion.CStone Pharmaceuticals and the National Key Research and Development Program of China.For the Chinese translation of the abstract see Supplementary Materials section.

Accumulated studies have discovered that circular RNAs (CircRNAs) are closely related to many complex human diseases. Due to this close relationship, CircRNAs can be used as good biomarkers for disease diagnosis and therapeutic targets for treatments. However, the number of experimentally verified circRNA-disease associations are still fewer and also conducting wet-lab experiments are constrained by the small scale and cost of time and labour. Therefore, effective computational methods are required to predict associations between circRNAs and diseases which will be promising candidates for small scale biological and clinical experiments. In this paper, we propose novel computational models based on Graph Convolution Networks (GCN) for the potential circRNA-disease association prediction. Currently most of the existing prediction methods use shallow learning algorithms. Instead, the proposed models combine the strengths of deep learning and graphs for the computation. First, they integrate multi-source similarity information into the association network. Next, models predict potential associations using graph convolution which explore this important relational knowledge of that network structure. Two circRNA-disease association prediction models, GCN based Node Classification (GCN-NC) and GCN based Link Prediction (GCN-LP) are introduced in this work and they demonstrate promising results in various experiments and outperforms other existing methods. Further, a case study proves that some of the predicted results of the novel computational models were confirmed by published literature and all top results could be verified using gene-gene interaction networks.

Hydrogen fuel cells are among the promising energy sources worldwide, which could accomplish cyclic production of energy and avoid the emission of green-house or contaminative byproducts. However, sulfur compounds (SCs) even at trace level (nmol/mol) are usually involved in cell construction and further H2 production, which would cause degradation of the catalysts and shorten the lifetime of the fuel cells. Moreover, the highly reactive SCs could cause varied species and concentrations of them in complex matrices, so online rather than offline analysis of SCs in H2 would be preferred. In this context, we developed a new system combining online cryogenic preconcentration of nine SCs and subsequent determination by GC-SCD (sulfur chemiluminescent detector), with the correlation coefficients of the calibration curves higher than 0.999, calculated limits of detection no higher than 0.050 nmol/mol, analytical time around 30 min per sample, and satisfactory precision and accuracy (RSD

Gastric cancer (GC) is one of the most common malignancies, with an ever-increasing incidence and high mortality rate. Chromobox4 (CBX4), also named hPC2, is a small ubiquitin-related modifier (SUMO) E3 ligase. Previous studies have found that high CBX4 expression is associated with tumor size, pathologic differentiation and decreased patient survival in hepatocellular carcinoma (HCC). However, the expression and prognostic value of CBX4 in GC have not been clarified. In our study, ONCOMINE, UALCAN, Kaplan-Meier Plotter, cBioPortal, DAVID 6.8 and TIMER were utilized. RT-PCR, immunohistochemistry (IHC), Western blot, CCK-8 assay, cell apoptosis assay, cell cycle assay were used to further verify in GC tissue samples or cell line. The transcriptional and protein level of CBX4 in GC tissues was found significantly elevated and a significant association between the expression of CBX4 and clinicopathological parameters was found in GC patients. Low expression of CBX4 in GC patients were correlated with a significantly improved prognosis. The functions of CBX4 are primarily related to the stem cell pluripotency signaling pathway, Hippo signaling pathway, HTLV-I infection, Notch signaling pathway, and N-glycan biosynthesis. Our results may provide novel insights for the selection of therapeutic targets and prognostic biomarkers for GC.

Background Despite multiple available fixed-dose combinations (FDCs) of inhaled long-acting β2-agonists (LABAs) plus long-acting muscarinic antagonists (LAMAs) and LABAs plus inhaled corticosteroids (ICS) for COPD, uncertainty remains regarding their comparative effects. Research Question Can comparative effectiveness and safety of LABA plus LAMA (LABA/LAMA) and LABA plus ICS (LABA/ICS) FDCs vary by different individual components of the dual combinations in COPD? Study Design and Methods We conducted a new user, propensity score-inverse probability of treatment weighting cohort study to compare the effectiveness and safety of two frequently used LABA/LAMA FDCs (indacaterol plus glycopyrronium [IND/GLY] and vilanterol plus umeclidinium [VI/UMEC]) vs three commonly prescribed LABA/ICS FDCs (salmeterol plus fluticasone propionate [SAL/FP], formoterol fumarate plus budesonide [FF/BUD], and formoterol fumarate plus beclomethasone dipropionate [FF/BDP]) using the Taiwanese nationwide health care claims from 2014 through 2017. The primary effectiveness outcome was the annual moderate to severe exacerbation rate, and safety outcomes included risks of severe pneumonia and cardiovascular disease requiring hospitalization. Weighted generalized linear mixed models and Cox proportional hazard models were used to assess the effectiveness and safety outcomes, respectively. Results Patients with COPD initiating IND/GLY and VI/UMEC showed an 11% (incidence rate ratio [IRR], 0.89; 95% CI, 0.80-0.98) and 20% (IRR, 0.80; 95% CI, 0.71-0.90) reduced annual rate of moderate to severe exacerbations, respectively, than those initiating SAL/FP, but showed a similar rate as those initiating FF/BUD or FF/BDP. Both LABA/LAMA FDCs, compared with SAL/FP and VI/UMEC vs FF/BDP, were associated with a 27% (hazard ratio [HR], 0.73; 95% CI, 0.59-0.90) to 42% (HR, 0.58; 95% CI, 0.48-0.70) reduced pneumonia risk. Cardiovascular risk was comparable in five groups. An intraclass difference existed in rates of moderate to severe COPD exacerbation and risks of pneumonia among LABA/ICS FDCs, but not between LABA/LAMA FDCs. Interpretation Both LABA/LAMAs vs SAL/FP are associated with a lower exacerbation rate and pneumonia risk, but exhibit similar effectiveness and safety outcomes compared with FF/BDP or FF/BUD, suggesting that comparative effects may differ by individual components of the dual therapies in COPD.

Na3(VO)2(PO4)2F (NVPOF) has been considered as one potential candidate for sodium-ion batteries because of its high operating voltage and theoretical capacity. However, the poor intrinsic electronic conductivity significantly restricts its widespread application. In response to this drawback, we adopt the optimization strategy of tuning the morphology and structure to boost the electrical conductivity and mitigate the capacity fading. In this paper, NVPOF microspheres with unique porous yolk-shell structure were fabricated via a facile one-step solvothermal method for the first time. By monitoring the morphological evolution with time-dependent experiments, the self-sacrifice and Ostwald ripening mechanism from rough spheres to yolk-shell structure was revealed. Benefited from the favorable interwoven nanosheets shell, inner cavity and porous core structure, the resulting NVPOF electrode exhibits superior rate capability of 63 mA h g−1 at 20 C as well as outstanding long-cycling performance with the capacity retention up to 92.1% over 1000 cycles at 5 C.

High-performance chemiresistor gas sensor made of sulfide porous GaN decorated with Pt nanoparticles, which shows tunable sensor response and enhanced sensitivity. The fabricated gas sensors show detection of H2 down to 30 ppm at 23 °C after sulfide treatment and Pt decorated porous GaN. The response time and recovery time were equal to 47 s and 113 s, respectively. Density functional theory simulations were used to support the detection mechanism based on sulfide treatment. Adsorption energy calculations showed that H adsorption energy is lowered by the simultaneous presence of S and Pt on the GaN (0001) surface. The density of states (DOS) calculations revealed possibility of bond strengthening when Pt and S is adsorbed on GaN surface along with H, arising from the hybridization of d and p orbitals of Pt and S with that of H 1s orbitals.

In this study, nitrilotriacetic acid (NTA) was used to enhance UV/Fe0/H2O2 process for highly salty wastewater at neutral pH. The addition of Fe0 could not accelerate the UV/H2O2 process for salty wastewater treatment at neutral pH due to the precipitation of Fe ions. While NTA added into UV/Fe0/H2O2 process which could form complex with Fe ion, the tartrazine removal could increase from 45.7% to 84.7% within 60 min. The k-value of the UV/Fe0-NTA/H2O2 system to remove tartrazine is also 2.2 times higher than that in UV/Fe0/H2O2 system. Meanwhile, the UV/Fe0-NTA/H2O2 process could obtain good performance for salt wastewater treatment under a wide salt concentration level of 0–0.2 M. It can also expand the application range of pH to neutral pH. SEM and XPS characterization showed that the corrosion rate of Fe0 in salty system was faster, and the presence of NTA could accelerate the recirculation of Fe3+/Fe2+ on the Fe0 surface. EPR spectroscopic analysis and scavenge experiments showed that •OH was the main reactive radical for the removal of tartrazine in the UV/Fe0-NTA/H2O2 system. In addition, the k-value for removing tartrazine in various types of salt systems in the UV/Fe0-NTA/H2O2 system was 3.2–4.2 times that of the UV/Fe0/H2O2 system. It demonstrated UV/Fe0-NTA/H2O2 processes fits to different types of salt system. Therefore, this study will provide a new method for the treatment of saline wastewater at neutral pH.

With the popularity of smart mobile equipment, the amount of data requested by users is growing rapidly. The traditional centralized processing method represented by the cloud computing model can no longer satisfy the effective processing of large amounts of data. Therefore, the mobile edge computing (MEC) is used as a new computing model to process the big growing data, which can better meet the service requirements. Similar to the task scheduling problem in cloud computing, an important issue in the MEC environment is task offloading and resource allocation. In this paper, we propose an adaptive task offloading and resource allocation algorithm in the MEC environment. The proposed algorithm uses the deep reinforcement learning (DRL) method to determine whether the task needs to be offloaded and allocates computing resources for the task. We simulate the generation of tasks in the form of Poisson distribution, and all tasks are submitted to be processed in the form of task flow. Besides, we consider the mobility of mobile user equipment (UE) between base stations (BSs), which is closer to the actual application environment. The DRL method is used to select the suitable computing node for each task according to the optimization objective, and the optimal strategy for solving the objective problem is learned in the algorithm training process. Compared with other comparison algorithms in different MEC environments, our proposed algorithm has the best performance in reducing the task average response time and the total system energy consumption, improving the system utility, which meets the profits of users and service providers.

The classification of mild cognitive impairment (MCI), which is a early stage of Alzheimer’s disease and is associated with brain structural and functional changes, is still a challenging task. Recent studies have shown great promise for improving the performance of MCI classification by combining multiple structural and functional features, such as grey matter volume and clustering coefficient. However, extracting which features and how to combine multiple features to improve the performance of MCI classification have always been challenging problems. To address these problems, in this study we propose a new method to enhance the feature representation of multi-modal MRI data by combining multi-view information to improve the performance of MCI classification. Firstly, we extract two structural features (including grey matter volume and cortical thickness) and two functional features (including clustering coefficient and shortest path length) of each cortical brain region based on automated anatomical labeling (AAL) atlas from both T1w MRI and rs-fMRI data of each subject. Then, in order to obtain features that are more helpful in distinguishing MCI subjects, an improved multi-task feature selection method, namely MTFS-gLASSO-TTR, is proposed. Finally, a multi-kernel learning algorithm is adopted to combine multiple features to perform the MCI classification task. Our proposed MCI classification method is evaluated on 315 subjects (including 105 LMCI subjects, 105 EMCI subjects and 105 NCs) with both T1w MRI and rs-fMRI data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Experimental results show that our proposed method achieves an accuracy of 88.5% and an area under the receiver operating characteristic (ROC) curve (AUC) of 0.897 for LMCI/NC classification, an accuracy of 82.7% and an AUC of 0.832 for EMCI/NC classification, and an accuracy of 79.6% and an AUC of 0.803 for LMCI/EMCI classification, respectively. In addition, by comparison, the accuracy and AUC values of our proposed method are better than those of some existing state-of-the-art methods in MCI classification. Overall, our proposed MCI classification method is effective and promising for automatic diagnosis of MCI in clinical practice.

This paper analyzes the existing zeroing neural network (ZNN) models from the perspective of control theory. It proposes an exclusive ZNN model for solving the dynamic complex-valued matrix Moore-Penrose inverse problem: the complex-valued zeroing neural network (CVZNN). Then, a method of constructing a special type of saturation-allowed activation function is defined, which relaxes the convex constraint on the activation function when constructing the ZNN model. The convergence of the CVZNN model activated by proposed saturation-allowed functions is analyzed. Besides, the robustness of the CVZNN model under different types of noise interference is investigated based on the perspective of the control theory. Finally, the effectiveness and superiority of the CVZNN model are illustrated by simulation experiments.

Prostate cancer is one of the leading causes of death in men worldwide, revealing a substantial heterogeneity in terms of molecular and clinical behaviors. Tumor infiltrating immune cell is associated with prognosis and response to immunotherapy in several cancer types. However, until now, the immune infiltrate profile of distinct subtypes for prostate cancer remains poorly characterized. In this study, using immune infiltration profiles as well as transcriptomic datasets, we characterized this subtype of prostate tumors. We observed that the FLI1 subtype of prostate tumors was highly enriched in immune system processes, immune related KEGG pathways and biological processes. We also expanded this approach to explore the immune infiltration profile of the high FLI1 expression subtype for skin cutaneous melanoma, similar results were found. Investigation of the association of immune infiltration features with the FLI1 expression demonstrated that many important features were associated with the FLI1 expression.

Pulmonary arterial hypertension (PAH) is a rare, but progressive and devastating vascular disease with few treatment options to prevent the advancement to right ventricular dysfunction hypertrophy and failure. Empagliflozin, a sodium-glucose cotransporter 2 (SGLT2) inhibitor, enhances urinary glucose excretion as well as reduces cardiovascular events and mortality in individuals with type 2 diabetes. While empagliflozin has been reported to lower systemic hypertension due to increased diuresis, the effect of empagliflozin on PAH is unknown. We used monocrotaline (MCT)-treated Sprague-Dawley rats to determine if empagliflozin alters PAH-associated outcomes. Compared to vehicle control, daily empagliflozin administration significantly improved survival in rats with severe MCT-induced PAH. Hemodynamic assessments showed that empagliflozin treatment significantly reduced mean pulmonary artery pressure, right ventricular systolic pressure, and increased pulmonary acceleration time. Empagliflozin treatment resulted in reduced right ventricular hypertrophy and fibrosis. Histological and molecular assessments of lung vasculature revealed significantly reduced medial wall thickening and decreased muscularization of pulmonary arterioles after empagliflozin treatment compared to vehicle-treated rats. In summary, SGLT2 inhibition with empagliflozin lowered mortality, reduced right ventricle systolic pressure, and attenuated maladaptive pulmonary remodeling in MCT-induced PAH. Clinical studies evaluating the efficacy of SGLT-2 inhibition should be considered for patients with PAH.