Sodium-ion batteries have received great attention because of the abundant sodium resources and low cost. As a typical kind of cathode materials for Na-ion batteries, sodium manganese oxides have shown great potential in cathode application due to their high specific capacity and good rate capability. Herein, we successfully synthesized P2-type Na 0.4 Mn 0.54 Co 0.46 O 2 nanosheets via a two-step annealing route.The morphology and structure information of Na 0.4 Mn 0.54 Co 0.46 O 2 products were characterized by X-ray diffraction (XRD), transmission electron microscope (TEM) and high resolution transmission electron microscope (HRTEM) technologies. The electrochemical performances of Na 0.4 Mn 0.54 Co 0.46 O 2 were measured by charge-discharge test, cyclic voltammogram (CV) and electrochemical impedance spectrum (EIS). As the cathode for Na-ion batteries, the layered Na 0.4 Mn 0.54 Co 0.46 O 2 nanosheets showed a high second charge capacity of 194 mAh/g and delivered a specific capacity of 125 mAh/g at a current of 20 mA/g after 60 cycles. 100 cycles at a rate of 2C) 30 . The available reversible capacity of P2-Na x [Fe 1/2 Mn 1/2 ]O 2 reaches 190 mAh/g with an average voltage of 2.75 V versus sodium metal 31 . The energy density is estimated to be 520 mWh/g, which is comparable to that of LiFePO 4 (about 530 mWh/g versus Li) and slightly higher than that of LiMn 2 O 4 (about 450mWh/g) 17,31 . The above research progress is proved that layered P2-type Co-doped sodium manganese oxides are promising cathode materials for Na-ion batteries. It is well know that reducing the manganese content and raising the average valence of manganese in the layered manganese-based cathode materials are effective ways to alleviate the manganese dissolution and Jahn-Teller effect. As we known, high sodium contents normally result in O3-type oxide cathodes, while low contents for P2-type ones with a higher capacity than O3-type 26,32,33 . Herein we have designed a layered Na 0.4 Mn 0.54 Co 0.46 O 2 cathode material with low sodium content for superior Na-ion batteries. The P2-Na 0.4 Mn 0.54 Co 0.46 O 2 have a good specific capacity and cycling performance at a current of 20 mA/g, and a specific capacity of 120 mAh/g is still achieved after 67 cycles. Experiment SectionMaterials synthesis: : : :MnCO 3 was synthesized by a precipitation method. In a typical synthesis, 10 mmol Mn(NO 3 ) 2 was dissolved in 200 mL distilled water, then 200 mL of 0.5 mol/L NH 4 HCO 3 was added into the Mn(NO 3 ) 2 solution. Spherical Mn 2 O 3 was synthesized by annealing microsphere MnCO 3 at 400 °C for 10 h in air condition. The P2-Na 0.4 Mn 0.54 Co 0.46 O 2 cathode was synthesized by mixing 5 mmol Mn 2 O 3 , 5 mmol uniformity of spherical morphology. The pure phase of Mn 2 O 3 precursor can be clearly proved by the corresponding XRD patterns (Figure S1 in Supplementary information). The morphology and size of P2-Na 0.4 Mn 0.54 Co 0.46 O 2 particles can be also clearly observed from SEM images (Figure 3c,d), revealing that during the following high-tem...
During the past decade, due to the number of proteins in PDB database being increased gradually, traditional methods cannot better understand the function of newly discovered enzymes in chemical reactions. Computational models and protein feature representation for predicting enzymatic function are more important. Most of existing methods for predicting enzymatic function have used protein geometric structure or protein sequence alone. In this paper, the functions of enzymes are predicted from many-sided biological information including sequence information and structure information. Firstly, we extract the mutation information from amino acids sequence by the position scoring matrix and express structure information with amino acids distance and angle. Then, we use histogram to show the extracted sequence and structural features respectively. Meanwhile, we establish a network model of three parallel Deep Convolutional Neural Networks (DCNN) to learn three features of enzyme for function prediction simultaneously, and the outputs are fused through two different architectures. Finally, The proposed model was investigated on a large dataset of 43,843 enzymes from the PDB and achieved 92.34% correct classification when sequence information is considered, demonstrating an improvement compared with the previous result.
Molecular visualization is often challenged with rendering of large molecular structures in real time. The key to LOD (level-of-detail), a classical technology, lies in designing a series of hierarchical abstractions of protein. In the paper, we improved the smoothness of transition for these abstractions by constructing a complete binary tree of a protein. In order to reduce the degree of expansion of the geometric model corresponding to the high level of abstraction, we introduced minimum ellipsoidal enveloping and some post-processing techniques. At the same time, a simple, ellipsoid drawing method based on graphics processing unit (GPU) is used that can guarantee that the drawing speed is not lower than the existing sphere-drawing method. Finally, we evaluated the rendering performance and effect on series of molecules with different scales. The post-processing techniques applied, diffuse shading and contours, further conceal the expansion problem and highlight the surface details.
RNA plays an important role in many biological processes, and RNA functions are primarily achieved by binding with a variety of proteins. But with the increasing complexity of RPIs networks, highthroughput biological techniques are usually expensive and time consuming. Therefore, there is an urgent need for high speed and reliably computational methods to predict RNA-protein interactions. In this study, we propose a hybrid deep learning model: RPI-MCNNBLSTM, which combines three convolutional neural networks (CNN) with a BLSTM network, to predict RNA-protein interactions using many-sided biological information including protein sequences, RNA sequence and structure. Firstly, we adopt a filling method to pad sequence and structure into equal length, and perform numerical encoding for the sequence and structure of the above equal length, respectively, which are appropriate for subsequent convolution operations. Secondly, we establish the three CNNs to learn the three biological information, separately, then use the BLSTM to capture the long range dependencies among the three features identified by the CNNs. The learned weighted representations are fed into a classification layer to predict ncRNA-protein interactions. Finally, the experimental results indicate that the proposed method achieves superior performance with the accuracy of 98.37% on the RPI1807 dataset, 92.99% on the RPI2241 dataset, 95.47% on the RPI369 dataset, 90.0% on the RPI448 and 87.4% on the RPI1446 dataset, respectively. The code of RPI-MCNNBLSTM and the datasets used in this work are available at https://github.com/xiaopang136/RPI for academic users.
Prediction of urban noise is becoming more significant for tackling noise pollution and protecting human mental health. However, the existing noise prediction algorithms neglected not only the correlation between noise regions, but also the nonlinearity and sparsity of the data, which resulted in low accuracy of filling in the missing entries of data. In this paper, we propose a model based on multiple views and kernel-matrix tensor decomposition to predict the noise situation at different times of day in each region. We first construct a kernel tensor decomposition model by using kernel mapping in order to speed decomposition rate and realize stable estimate the prediction system. Then, we analyze and compute the cause of the noise from multiple views including computing the similarity of regions and the correlation between noise categories by kernel distance, which improves the credibility to infer the noise situation and the categories of regions. Finally, we devise a prediction algorithm based on the kernel-matrix tensor factorization model. We evaluate our method with a real dataset, and the experiments to verify the advantages of our method compared with other existing baselines.
Recurrent geometric network(RGN) as a deep learning model have been successfully applied to predict protein 3D structure and achieved better results than conventional methods. However, because of the inter- nal complexity and nonlinear structure of deep neural networks, the underlying processes why the model achieving such performance is challenging and sometimes difficult to explain. In this paper, for pro- tein 3D structure prediction tasks, we analyze these hidden states of recurrent geometric network from a new perspective. Firstly, a method is proposed to explain network position characteristic by searching for the most similar amino acids near the selected amino acids net- work. We also analyze the update response of new neurons in forward and backward networks to further focus on the position characteris- tics of hidden state. Then, a superimposed visualization method is put forward to study which and when neurons in bidirectional net- works have a greater influence on the prediction results. Finally, we adopt comparison method for torsion angles to analyze how the back- ward neuron affects the final prediction results. The usability and effectiveness of our method are demonstrated through case studies.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.