Prediction of drug-target interactions (DTI) plays a vital role in drug development in various areas, such as virtual screening, drug repurposing and identification of potential drug side effects. ...Despite extensive efforts have been invested in perfecting DTI prediction, existing methods still suffer from the high sparsity of DTI datasets and the cold start problem. Here, we develop KGE_NFM, a unified framework for DTI prediction by combining knowledge graph (KG) and recommendation system. This framework firstly learns a low-dimensional representation for various entities in the KG, and then integrates the multimodal information via neural factorization machine (NFM). KGE_NFM is evaluated under three realistic scenarios, and achieves accurate and robust predictions on four benchmark datasets, especially in the scenario of the cold start for proteins. Our results indicate that KGE_NFM provides valuable insight to integrate KG and recommendation system-based techniques into a unified framework for novel DTI discovery.
Graph neural networks (GNN) has been considered as an attractive modelling method for molecular property prediction, and numerous studies have shown that GNN could yield more promising results than ...traditional descriptor-based methods. In this study, based on 11 public datasets covering various property endpoints, the predictive capacity and computational efficiency of the prediction models developed by eight machine learning (ML) algorithms, including four descriptor-based models (SVM, XGBoost, RF and DNN) and four graph-based models (GCN, GAT, MPNN and Attentive FP), were extensively tested and compared. The results demonstrate that on average the descriptor-based models outperform the graph-based models in terms of prediction accuracy and computational efficiency. SVM generally achieves the best predictions for the regression tasks. Both RF and XGBoost can achieve reliable predictions for the classification tasks, and some of the graph-based models, such as Attentive FP and GCN, can yield outstanding performance for a fraction of larger or multi-task datasets. In terms of computational cost, XGBoost and RF are the two most efficient algorithms and only need a few seconds to train a model even for a large dataset. The model interpretations by the SHAP method can effectively explore the established domain knowledge for the descriptor-based models. Finally, we explored use of these models for virtual screening (VS) towards HIV and demonstrated that different ML algorithms offer diverse VS profiles. All in all, we believe that the off-the-shelf descriptor-based models still can be directly employed to accurately predict various chemical endpoints with excellent computability and interpretability.
Display omitted
•MFCs and a CDI cell were integrated for domestic wastewater treatment.•CDI was powered by two MFCs in series and parallel.•Simultaneous removal organic pollutants and subsequent ...desalination was achieved.•The energy generated by MFCs was used to remove ions in the MFC effluent with CDI.•Nutrient removal was achieved by biological processes in MFCs and electrosorption in CDI.
In the microbial fuel cells (MFCs) driven capacitive deionization (CDI) process, the electricity generated from MFCs can be utilized for downstream deionization processes in the CDI cell. However, due to the typical configuration of MFCs, i.e., anaerobic anode and aerobic cathode compartments, the MFC effluent still contains various ionic pollutants, e.g., NO3− and PO43−. In this study, we present the proof-of-concept of an integrated MFC-CDI process for advanced wastewater treatment: the cost-effective continuous-flow MFCs perform as secondary treatment for organic carbon and ammonium removal, electricity production; the energy-efficient CDI cell works as the tertiary treatment to enhance the purification and desalination of domestic wastewater effluent. The results indicated that approximately 90% of the chemical oxygen demand (COD) and ammonium were removed in the MFCs. CDI, powered by connecting two MFCs in parallel and associated with a potential of 0.49V, could subsequently remove multi-ionic species from the MFC-treated effluent. More specifically, nutrient ions (NO3− and PO43−) can be electrostatically removed by the CDI process. These results demonstrated that the integrated MFC-CDI technology, which harvested energy from wastewater, held great potential to be an advanced, energy-saving purification process for the simultaneous removal of carbonaceous pollutants, improved nutrient removal, and further electrosorptive desalination of domestic wastewater treatment, in order to achieve a sustainable water-energy system. They also showed that, in addition to effluent quality improvement, the tertiary electrosorption of CDI played a significant role in the stability of effluent quality so as to dampen variations in the quality of secondary-treated effluent.
Apoptosis plays a dual role in cancer development and malignancy. The role of apoptosis-related caspases in cancer remains controversial, particularly in oral tongue squamous cell carcinoma (OTSCC). ...In this study, we examined the protein levels of cleaved caspase-3, caspase-3, caspase-8, and caspase-9 on tissue microarrays consisting of samples from 246 OTSCC patients by immunohistochemistry. Wilcoxon signed-rank test indicated that the protein levels of cleaved caspase-3, caspase-3, caspase-8, and caspase-9 in tumor tissues were significantly higher compared to those in adjacent normal tissues (all p<0.001). The expression level of caspase-8 in tumors was elevated in patients with lymph node invasion. Moreover, positive expression of cleaved caspase-3 was associated with shorter disease-free survival (DFS) in OTSCC patients with moderate differentiation and lymph node invasion. Combination of either positive cleaved caspase-3 or higher caspase-3 expression or both was associated with poor DFS. Interestingly, stratification analysis showed that co-expression levels of positive cleaved caspase-3 or/and higher caspase-3 were associated with better disease-specific survival in patients with advanced stages of the disease, such as large tumor size and lymph node invasion, whereas it was associated with poor DFS in OTSCC patients with moderate cell differentiation and small tumor size. Taken together, cleaved caspase-3 and caspase-3/8/9 could be biomarkers for tumorigenesis in OTSCC patients. The co-expression level of cleaved caspase-3 and caspase-3 might be a prognostic biomarker for OTSCC patients, particular in those patients with certain tumor stages and cell differentiation status.
Electron fractionalization is intimately related to topology. In one-dimensional systems, fractionally charged states exist at domain walls between degenerate vacua. In two-dimensional systems, ...fractionalization exists in quantum Hall fluids, where time-reversal symmetry is broken by a large external magnetic field. Recently, there has been a tremendous effort in the search for examples of fractionalization in two-dimensional systems with time-reversal symmetry. In this Letter, we show that fractionally charged topological excitations exist on graphenelike structures, where quasiparticles are described by two flavors of Dirac fermions and time-reversal symmetry is respected. The topological zero modes are mathematically similar to fractional vortices in p-wave superconductors. They correspond to a twist in the phase in the mass of the Dirac fermions, akin to cosmic strings in particle physics.
This study introduces a recent field experiment investigating multiscale terrain–circulation–precipitation interactions. When a synoptic‐scale northeasterly wind prevails under the active East Asian ...winter monsoon, stratocumulus cloud decks with severe rainfall exceeding 100 mm·day−1 frequently occur in the northeastern plain area and adjacent mountains in Yilan, Taiwan. The Yilan Experiment of Severe Rainfall (YESR2020) is a field campaign from November 20, 2020, to November 24, 2020, to survey the physical processes leading to severe wintertime rainfall. The three‐dimensional structure of the wind field and the atmospheric environment can be identified through high temporal and spatial resolution sounding observations, which is empowered by the novel Storm Tracker mini‐radiosonde. During YESR2020, the continuously collected meteorological data of two northeasterly episodes captured the variability of local‐scale wind patterns and the features of the severe rainfall induced by stratocumulus. A preliminary analysis indicated that a local‐scale convergence line could appear over the plain area of Yilan under the northeasterly environmental condition. The precipitation hotspot was located in the mountain region of southern Yilan, where the local winds signified turbulence features. Moreover, the severe rainfall of the two northeasterly episodes spotlighted shallow cumulus under stratus with pure warm rain processes. The results of YESR2020 inspire the arrangement of future field observations to explore detailed mechanisms of heavily precipitating stratocumulus over complex topography.
We conducted the Yilan Experiment of Severe Rainfall (YESR2020) to survey physical processes leading to severe rainfall in the northeastern plain area and adjacent mountains in Yilan, Taiwan, when a synoptic‐scale northeasterly wind prevails under the active East Asian winter monsoon with stratocumulus cloud decks. A preliminary analysis indicated that a local‐scale convergence line appeared over the plain area, and the precipitation hotspot was located in the mountain region of southern Yilan, where turbulence features were apparent. The results inspire the arrangement of future field observations to explore mechanisms of heavily precipitating stratocumulus over complex topography.
Summary
Although the health of rivers is threatened by multiple anthropogenic stressors with increasing frequency, it remains an open question how riverine microbial communities respond to emerging ...micropollutants. Here, by using 16S rDNA amplicon sequencing of 60 water samples collected during different hydrological seasons, we investigated the spatio‐temporal variation and the co‐occurrence patterns of microbial communities in the anthropogenically impacted Jiulong River in China. The results indicated that the riverine microbial co‐occurrence network had a nonrandom, modular structure, which was mainly shaped by the taxonomic relatedness of co‐occurring species. Fecal indicator bacteria may survive for prolonged periods of time in river water, but they formed an independent module which had fewer interactions with typical freshwater bacteria. Multivariate analysis demonstrated that nutrients and micropollutants i.e., pharmaceuticals and personal care products (PPCPs) exerted combined effects in shaping α‐ and β‐diversity of riverine microbial communities. Remarkably, we showed that a hitherto unrecognized disruptive effect of PPCPs on the abundance variations of central species and module communities was stronger than the influence of physicochemical factors, suggesting the key role played by micropollutants for the microbial co‐occurrence relationships in lotic ecosystems. Overall, our findings provide novel insights into community assembly in aquatic environments experiencing anthropogenic stresses.
Based on the sense of place theory and the design principles of guidance and interpretation, this study developed an augmented reality mobile guidance system that used a historical ...geo-context-embedded visiting strategy. This tool for heritage guidance and educational activities enhanced visitor sense of place. This study consisted of 3 visitor groups (i.e., AR-guidance, audio-guidance, and no-guidance) composed of 87 university students. A quasi-experimental design was adopted to evaluate whether augmented reality guidance more effectively promoted sense of place and learning performance than the other groups. The results indicated that visitors who used AR guidance showed significant learning and sense of place effects. Interviews were also employed to determine the possible factors that contribute to the formation of sense of place. Finally, a majority of the visitors who participated in the study demonstrated positive attitudes toward the use of the AR-guidance system.
Ligand docking (LD), a technology for predicting protein-ligand (PL)-binding conformations and strengths, plays key roles in virtual screening (VS). However, the accuracy and speed of current LD ...methodologies remain suboptimal. Here, we discuss how deep learning (DL) could help to bridge this gap by examining recent advancements and projecting future trends.
Safety is a main reason for drug failures, and therefore, the detection of compound toxicity and potential adverse effects in the early stage of drug development is highly desirable. However, ...accurate prediction of many toxicity endpoints is extremely challenging due to low accessibility of sufficient and reliable toxicity data, as well as complicated and diversified mechanisms related to toxicity. In this study, we proposed the novel multitask graph attention (MGA) framework to learn the regression and classification tasks simultaneously. MGA has shown excellent predictive power on 33 toxicity data sets and has the capability to extract general toxicity features and generate customized toxicity fingerprints. In addition, MGA provides a new way to detect structural alerts and discover the relationship between different toxicity tasks, which will be quite helpful to mine toxicity information from large amounts of toxicity data.