Search for: All records

Improving the fireproof performance of polymers is crucial for ensuring human safety and enabling future space colonization. However, the complexity of the mechanisms for flame retardant and the need for customized material design pose significant challenges. To address these issues, we propose a machine learning (ML) framework based on substructure fingerprinting and self-enforcing deep neural networks (SDNN) to predict the fireproof performance of flame-retardant epoxy resins. Our model is based on a comprehensive understanding of the physical mechanisms of materials and can predict fireproof performance and eliminate the needs for properties descriptors, making it more convenient than previous ML models. With a dataset of only 163 samples, our SDNN models show an average prediction error of 3% for the limited oxygen index (LOI). They also provide satisfactory predictions for the peak of heat release rate PHR and total heat release (THR), with coefficient of determination (R2) values of 0.87 and 0.85, respectively, and average prediction errors less than 17%. Our model outperforms the support vector model SVM for all three indices, making it a state-of-the-art study in the field of flame retardancy. We believe that our framework will be a valuable tool for the design and virtual screening of flame retardants and will contribute to the development of safer and more efficient polymer materials. 

Spatially resolved scRNA-seq (sp-scRNA-seq) technologies provide the potential to comprehensively profile gene expression patterns in tissue context. However, the development of computational methods lags behind the advances in these technologies, which limits the fulfillment of their potential. In this study, we develop a deep learning approach for clustering sp-scRNA-seq data, named Deep Spatially constrained Single-cell Clustering (DSSC). In this model, we integrate the spatial information of cells into the clustering process in two steps: (1) the spatial information is encoded by using a graphical neural network model, and (2) cell-to-cell constraints are built based on the spatial expression pattern of the marker genes and added in the model to guide the clustering process. Then, a deep embedding clustering is performed on the bottleneck layer of autoencoder by Kullback–Leibler (KL) divergence along with the learning of feature representation. DSSC is the first model that can use information from both spatial coordinates and marker genes to guide cell/spot clustering. Extensive experiments on both simulated and real data sets show that DSSC boosts clustering performance significantly compared with the state-of-the-art methods. It has robust performance across different data sets with various cell type/tissue organization and/or cell type/tissue spatial dependency. We conclude that DSSC is a promising tool for clustering sp-scRNA-seq data. 

Due to its fast switching speed, the voltage sharing of series-connected SiC MOSFETs is more sensitive to the parasitic components from the power modules and the system, which results in more challenges for voltage balancing control. For two series-connected SiC MOSFETs realized by one half-bridge module, the detailed analysis and measurement indicate that the unbalanced parasitic capacitors inside the power module comprise the dominant factor causing the difference of turn-off dv/dt. In this paper, the traditional gate turn-off delay-time control is first used as an example to analyze the limitation of the existing active voltage balancing (AVB) control methods under AC load current: 1) AVB control has a limitation to adjust delay time accurately under AC current; 2) the voltage imbalance of the body diodes cannot be solved by AVB control. To achieve voltage balancing control of series-connected SiC MOSFETs and body diodes, this paper proposes a new two-part hybrid approach: 1) passive dv/dt compensation: one small compensation capacitor is applied to balance the non-uniform distribution of parasitic capacitors inside the power module, so the series-connected MOSFETs can have the same turn-off dv/dt; 2) active gate signal turn-off time adjustment: a closed-loop delay time control is applied to compensate the gate signal mismatch of MOSFETs. To verify the proposed balancing approach, a single-phase pump-back test is conducted to show the improvement of voltage sharing of both MOSFETs and body diodes. 

ABSTRACT Although Wnt signaling is clearly important for the intestinal epithelial homeostasis, the relevance of various sources of Wnt ligands themselves remains incompletely understood. Blocking the release of Wnt in distinct stromal cell types suggests obligatory functions of several stromal cell sources and yields different observations. The physiological contribution of epithelial Wnt to tissue homeostasis remains unclear. We show here that blocking epithelial Wnts affects colonic Reg4+ epithelial cell differentiation and impairs colonic epithelial regeneration after injury in mice. Single-cell RNA analysis of intestinal stroma showed that the majority of Wnt-producing cells were contained in transgelin (Tagln+) and smooth muscle actin α2 (Acta2+) expressing populations. We genetically attenuated Wnt production from these stromal cells using Tagln-Cre and Acta2-CreER drivers, and found that blockage of Wnt release from either epithelium or Tagln+ and Acta2+ stromal cells impaired colonic epithelial healing after chemical-induced injury. Aggregated blockage of Wnt release from both epithelium and Tagln+ or Acta2+ stromal cells drastically diminished epithelial repair, increasing morbidity and mortality. These results from two uncharacterized stromal populations suggested that colonic recovery from colitis-like injury depends on multiple Wnt-producing sources.