UNet based on dynamic convolution decomposition and triplet attention.

The robustness and generalization of medical image segmentation models are being challenged by the differences between different disease types, different image types, and different cases.Deep learning based semantic segmentation methods have been providing state-of-the-art performance in the last few years. One deep learning technique, U-Net, has become the most popular architecture in the medical imaging segmentation. Despite outstanding overall performance in segmenting medical images, it still has the problems of limited feature expression ability and inaccurate segmentation. To this end, we propose a DTA-UNet based on Dynamic Convolution Decomposition (DCD) and Triple Attention (TA). Firstly, the model with Attention U-Net as the baseline network uses DCD to replace all the conventional convolution in the encoding-decoding process to enhance its feature extraction capability. Secondly, we combine TA with Attention Gate (AG) to be used for skip connection in order to highlight lesion regions by removing redundant information in both spatial and channel dimensions. The proposed model are tested on the two public datasets and actual clinical dataset such as the public COVID-SemiSeg dataset, the ISIC 2018 dataset, and the cooperative hospital stroke segmentation dataset. Ablation experiments on the clinical stroke segmentation dataset show the effectiveness of DCD and TA with only a 0.7628 M increase in the number of parameters compared to the baseline model. The proposed DTA-UNet is further evaluated on the three datasets of different types of images to verify its universality. Extensive experimental results show superior performance on different segmentation metrics compared to eight state-of-art methods.The GitHub URL of our code is https://github.com/shuaihou1234/DTA-UNet .

Study on eco-efficiency of industrial parks in China based on data envelopment analysis.

China's industrial parks have been playing a crucial role on driving regional economy development, but also been posing threats to local environment due to intensive resource consumption and waste emission. Chinese government facilitated eco-industrial development of industrial park, aiming to output more with less environmental burden. In our study, the eco-efficiency levels of 40 Chinese industrial parks in 2012 were assessed and ranked by Data Envelopment Analysis (DEA). This paper applied indicators relevant to resource, economy, and environment from industrial parks which can well reflect the characteristics of eco-efficiency conforming to the concept of sustainability. This paper introduced how to adjust less sustainable parks to be more sustainable according to the DEA results. The roles of industrial added value per capita, industrial structure, environmental policy and development scale as influence factors of eco-efficiency were discussed. The results show that large differences exist in the eco-efficiency of different industrial parks. It is shown that 20% of the parks are relatively efficient. 47% of the study parks being inefficient in terms of scale efficiency show decreasing returns to scale. Policy implementations for the management of industrial parks were also discussed based on the results.

O-GlcNAc regulates NEDD4-1 stability via caspase-mediated pathway.

O-GlcNAc modification of cytosolic and nuclear proteins regulates essential cellular processes such as stress responses, transcription, translation, and protein degradation. Emerging evidence indicates O-GlcNAcylation has a dynamic interplay with ubiquitination in cellular regulation. Here, we report that O-GlcNAc indirectly targets a vital E3 ubiquitin ligase enzyme of NEDD4-1. The protein level of NEDD4-1 is accordingly decreased following an increase of overall O-GlcNAc level upon PUGNAc or glucosamine stimulation. O-GlcNAc transferase (OGT) knockdown, overexpression and mutation results confirm that the stability of NEDD4-1 is negatively regulated by cellular O-GlcNAc. Moreover, the NEDD4-1 degradation induced by PUGNAc or GlcN is significantly inhibited by the caspase inhibitor. Our study reveals a regulation mechanism of NEDD4-1 stability by O-GlcNAcylation.

Proteomic analysis of O-GlcNAcylated proteins in invasive ductal breast carcinomas with and without lymph node metastasis.

The potential role of protein O-GlcNAcylation in cancer has been studied extensively, and the spread of cancer cells to regional lymph nodes is the first step in the dissemination of breast cancer. However, the correlation between O-GlcNAcylation and lymphatic metastasis in breast cancer remains elusive. In this study, we demonstrated that the overall O-GlcNAcylation as well as O-GlcNAc transferase (OGT) tends to decrease in response to the augmentation of lymph node metastasis (LNM) in invasive ductal breast carcinomas (IDCs). Although accumulating evidence indicates that individual O-GlcNAcylation may be important in the pathogenesis of breast cancer, O-GlcNAcylated proteins in IDCs are still largely unexplored. Herein, O-GlcNAcylated proteins of IDCs were chemo-enzymatically enriched and identified via liquid chromatography combined with tandem mass spectrometry. In total, 155 O-GlcNAcylated proteins were determined, of which 41 were only observed in LNM tissues, while 40 were unique in non-LNM samples. Gene ontology analysis showed that O-GlcNAc is primarily a nucleocytoplasmic post-translational modification, and most enriched functional terms were related to cancer development in both metastatic and non-metastatic IDCs. Moreover, several O-GlcNAcylated proteins involved in glycolysis and its accessory pathway were identified from LNM and non-LNM groups, respectively. These results indicate that the O-GlcNAcylation statuses of individual proteins were independent of the overall O-GlcNAcylation levels of metastatic and non-metastatic IDCs. Aberrant O-GlcNAc modification of these proteins might be associated with LNM progression.

Three-Dimensional Ordered Mesoporous MnO2-Supported Ag Nanoparticles for Catalytic Removal of Formaldehyde.

Three-dimensional (3D) ordered mesoporous Ag/MnO2 catalyst was prepared by impregnation method based on 3D-MnO2 and used for catalytic oxidation of HCHO. Ag nanoparticles are uniformly distributed on the polycrystalline wall of 3D-MnO2. The addition of Ag does not change the 3D ordered mesoporous structure of the Ag/MnO2, but does reduce the pore size and surface area. Ag nanoparticles provide sufficient active site for the oxidation reaction of HCHO, and Ag (111) crystal facets in the Ag/MnO2 are active faces. The 8.9% Ag/MnO2 catalyst shows a higher normalized rate (10.1 nmol·s(-1)·m(-2) at 110 °C) and TOF (0.007 s(-1) at 110 °C) under 1300 ppm of HCHO and 150 000 h(-1) of GHSV, and its apparent activation energy of the reaction is the lowest (39.1 kJ/mol). More Ag active sites, higher low-temperature reducibility, more abundant surface lattice oxygen species, oxygen vacancies, and lattice defects generated from interaction Ag with MnO2 are responsible for the excellent catalytic performance of HCHO oxidation on the 8.9% Ag/MnO2 catalyst. The 8.9% Ag/MnO2 catalyst remained highly active and stable under space velocity increasing from 60 000 to 150 000 h(-1), under initial HCHO concentration increasing from 500 to 1300 ppm, and under the presence of humidity, respectively.