ScLSTM: single-cell type detection by siamese recurrent network and hierarchical clustering.

MOTIVATION: Categorizing cells into distinct types can shed light on biological tissue functions and interactions, and uncover specific mechanisms under pathological conditions. Since gene expression throughout a population of cells is averaged out by conventional sequencing techniques, it is challenging to distinguish between different cell types. The accumulation of single-cell RNA sequencing (scRNA-seq) data provides the foundation for a more precise classification of cell types. It is crucial building a high-accuracy clustering approach to categorize cell types since the imbalance of cell types and differences in the distribution of scRNA-seq data affect single-cell clustering and visualization outcomes. RESULT: To achieve single-cell type detection, we propose a meta-learning-based single-cell clustering model called ScLSTM. Specifically, ScLSTM transforms the single-cell type detection problem into a hierarchical classification problem based on feature extraction by the siamese long-short term memory (LSTM) network. The similarity matrix derived from the improved sigmoid kernel is mapped to the siamese LSTM feature space to analyze the differences between cells. ScLSTM demonstrated superior classification performance on 8 scRNA-seq data sets of different platforms, species, and tissues. Further quantitative analysis and visualization of the human breast cancer data set validated the superiority and capability of ScLSTM in recognizing cell types.

High-Photoresponsivity Self-Powered a-, ε-, and ß-Ga2O3/p-GaN Heterojunction UV Photodetectors with an In Situ GaON Layer by MOCVD.

In this paper, self-powered ultraviolet (UV) photodetectors with high response performance based on Ga2O3/p-GaN were fabricated by metal-organic chemical vapor deposition (MOCVD). The effects of different crystal phases of Ga2O3 (including a, ε, ε/ß, and ß) grown on p-GaN films on the performance of photodetectors were systematically studied. Moreover, an in situ GaON dielectric layer improved the responsivity of Ga2O3/p-GaN photodetectors by 20 times. All Ga2O3/p-GaN photodetectors showed self-power capability without bias. An ultralow dark current of 3.08 pA and a Iphoto/Idark ratio of 4.1 × 103 (1.8 × 103) under 254 nm (365 nm) light were obtained for the ß-Ga2O3/p-GaN photodetector at 0 V bias. Furthermore, the ß-Ga2O3/p-GaN photodetector showed excellent sensitivity with a high responsivity of 3.8 A/W (0.83 A/W), a fast response speed of 66/36 ms (36/73 ms), and a high detectivity of 1.12 × 1014 Jones (2.44 × 1013 Jones) under 254 nm (365 nm) light at 0 V bias. The carrier transport mechanism of the Ga2O3/p-GaN self-powered photodetector was also analyzed through the device energy band diagram. This work provides critical information for the design and fabrication of high-performance self-powered Ga2O3/p-GaN UV photodetectors, opening the door to a variety of photonic systems and applications without an external power supply.

Thermo-Responsive Quaternary Ammonium Chitosan Nanocapsules with On-Demand Controlled Pesticide Release and Maximally Synergistic Biological Activity.

Pesticides play an important role in pest control. However, they can be limited due to low utilization efficiency, causing substantial losses to the environment and ecological damage. Nanotechnology is an active area of research regarding encapsulation of pesticides for sustainable pest control. Here, we developed intelligent formulations of avermectin (Av) quaternary ammonium chitosan surfactant (QACS) nanocapsules (i.e., Av-Th@QACS) with on-demand controlled release properties, toward ambient temperature and maximal synergistic biological activity of Av and QACS. The Av-Th@QACS regulated the quantity of pesticide release in accordance with the ambient temperature changes and, insofar as this release is a means of responding to variations in pest populations, maximized the synergistic activity. In addition, the Av-Th@QACS were highly adhesive to crop leaves as a result of the prolonged retention time on the crop leaves. Therefore, Av-Th@QACS exhibited greater control against aphids at 35 °C than at 15 and 25 °C. Compared with commercial formulations, Av-Th@QACS was more toxic at 35 °C and less toxic at 15 °C. Thus, researchers can apply Av-Th@QACS as intelligent nanopesticides with an on-demand, controlled release and synergistic biological activity and, in so doing, prolong pesticide duration and improve the utilization efficiency.

Highly heterogeneous epitaxy of flexoelectric BaTiO3-δ membrane on Ge.

The integration of complex oxides with a wide spectrum of functionalities on Si, Ge and flexible substrates is highly demanded for functional devices in information technology. We demonstrate the remote epitaxy of BaTiO3 (BTO) on Ge using a graphene intermediate layer, which forms a prototype of highly heterogeneous epitaxial systems. The Ge surface orientation dictates the outcome of remote epitaxy. Single crystalline epitaxial BTO3-δ films were grown on graphene/Ge (011), whereas graphene/Ge (001) led to textured films. The graphene plays an important role in surface passivation. The remote epitaxial deposition of BTO3-δ follows the Volmer-Weber growth mode, with the strain being partially relaxed at the very beginning of the growth. Such BTO3-δ films can be easily exfoliated and transferred to arbitrary substrates like Si and flexible polyimide. The transferred BTO3-δ films possess enhanced flexoelectric properties with a gauge factor of as high as 1127. These results not only expand the understanding of heteroepitaxy, but also open a pathway for the applications of devices based on complex oxides.

Development of Chlorantraniliprole and Lambda Cyhalothrin Double-Loaded Nano-Microcapsules for Synergistical Pest Control.

Nanotechnology could greatly improve global agricultural food production. Chlorantraniliprole and lambda cyhalothrin double-loaded nano-microcapsules were fabricated to enhance the control of pests by pesticides and improve the pesticide utilization efficiency. The nano-microcapsules were synthesized using a method involving the solid in oil in water encapsulation technique and solvent evaporation. The nano-microcapsules slowly and simultaneously released lambda cyhalothrin and chlorantraniliprole. The cumulative lambda cyhalothrin and chlorantraniliprole release rates at 40 h were 80% and 70%, respectively. Indoor Spodoptera frugiperda control tests indicated that the double-loaded nano-microcapsules were more toxic than lambda cyhalothrin water-dispersible granules, chlorantraniliprole water-dispersible granules, and a mixture of lambda cyhalothrin water-dispersible granules and chlorantraniliprole water-dispersible granules, indicating that the pesticides in the nano-microcapsules synergistically controlled Spodoptera frugiperda. The results indicated that pesticide nano-microcapsules with synergistic effects can be developed that can improve the effective pesticide utilization efficiency and pesticide bioavailability. This is a new idea for achieving environmentally intelligent pesticide delivery.

Domain-adversarial multi-task framework for novel therapeutic property prediction of compounds.

MOTIVATION: With the rapid development of high-throughput technologies, parallel acquisition of large-scale drug-informatics data provides significant opportunities to improve pharmaceutical research and development. One important application is the purpose prediction of small-molecule compounds with the objective of specifying the therapeutic properties of extensive purpose-unknown compounds and repurposing the novel therapeutic properties of FDA-approved drugs. Such a problem is extremely challenging because compound attributes include heterogeneous data with various feature patterns, such as drug fingerprints, drug physicochemical properties and drug perturbation gene expressions. Moreover, there is a complex non-linear dependency among heterogeneous data. In this study, we propose a novel domain-adversarial multi-task framework for integrating shared knowledge from multiple domains. The framework first uses an adversarial strategy to learn target representations and then models non-linear dependency among several domains. RESULTS: Experiments on two real-world datasets illustrate that our approach achieves an obvious improvement over competitive baselines. The novel therapeutic properties of purpose-unknown compounds that we predicted have been widely reported or brought to clinics. Furthermore, our framework can integrate various attributes beyond the three domains examined herein and can be applied in industry for screening significant numbers of small-molecule drug candidates. AVAILABILITY AND IMPLEMENTATION: The source code and datasets are available at https://github.com/JohnnyY8/DAMT-Model. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.