Convolution Filter Compression via Sparse Linear Combinations of Quantized Basis.

Convolutional neural networks (CNNs) have achieved significant performance on various real-life tasks. However, the large number of parameters in convolutional layers requires huge storage and computation resources, making it challenging to deploy CNNs on memory-constrained embedded devices. In this article, we propose a novel compression method that generates the convolution filters in each layer using a set of learnable low-dimensional quantized filter bases. The proposed method reconstructs the convolution filters by stacking the linear combinations of these filter bases. By using quantized values in weights, the compact filters can be represented using fewer bits so that the network can be highly compressed. Furthermore, we explore the sparsity of coefficients through L1 -ball projection when conducting linear combination to further reduce the storage consumption and prevent overfitting. We also provide a detailed analysis of the compression performance of the proposed method. Evaluations of image classification and object detection tasks using various network structures demonstrate that the proposed method achieves a higher compression ratio with comparable accuracy compared with the existing state-of-the-art filter decomposition and network quantization methods.

Efficient photoanode with a MoS2/TiO2/Au nanoparticle heterostructure for ultraviolet-visible photoelectrocatalysis.

Green energy technology is generally becoming one of hot issues that need to be solved due to the adverse effects on the environment of fossil fuels. One of the strategies being studied and developed by theorists and experimentalists is the use of photoelectrochemical (PEC) cells, which are emerging as a candidate to produce hydrogen from water splitting. However, creating photoelectrodes that meet the requirements for PEC water splitting has emerged as the primary obstacle in bringing this technology to commercial fruition. Here, we construct a heterostructure, which consists of MoS2/TiO2/Au nanoparticles (NPs) to overcome the drawbacks of the photoanode. Owing to the dependence on charge transfer, the bandgap of MoS2/TiO2and the utilization the Au NPs as a stimulant for charges separation of TiO2by localized surface plasmon resonances effect as well as the increase of hot electron injection to cathode, leading to photocatalytic activities are improved. The results have recorded a significant increase in the photocurrent density from 2.3µAcm-2of TiO2to approximately 16.3µAcm-2of MoS2/TiO2/Au NPs. This work unveils a promising route to enhance the visible light adsorption and charge transfer in photo-electrode of the PEC cells by combining two-dimensional materials with metal NPs.

Ultrasound-guided stellate ganglion block benefits the postoperative recovery of patients undergoing laparoscopic colorectal surgery: a single-center, double-blinded, randomized controlled clinical trial.

BACKGROUND: With the increasing prevalence of colorectal cancer (CRC), optimizing perioperative management is of paramount importance. This study investigates the potential of stellate ganglion block (SGB), known for its stress response-mediating effects, in improving postoperative recovery. We postulate that preoperative SGB may enhance the postoperative recovery of patients undergoing laparoscopic CRC surgery. METHODS: We conducted a randomized controlled trial of 57 patients undergoing laparoscopic colorectal cancer surgery at a single center. Patients, aged 18-70 years, were randomly assigned to receive either preoperative SGB or standard care. SGB group patients received 10 mL of 0.2% ropivacaine under ultrasound guidance prior to surgery. Primary outcome was time to flatus, with secondary outcomes encompassing time to defecation, lying in bed time, visual analog scale (VAS) pain score, hospital stays, patient costs, intraoperative and postoperative complications, and 3-year mortality. A per-protocol analysis was used. RESULTS: Twenty-nine patients in the SGB group and 28 patients in the control group were analyzed. The SGB group exhibited a significantly shorter time to flatus (mean [SD] hour, 20.52 [9.18] vs. 27.93 [11.69]; p = 0.012), accompanied by decreased plasma cortisol levels (mean [SD], postoperatively, 4.01 [3.42] vs 7.75 [3.13], p = 0.02). Notably, postoperative pain was effectively managed, evident by lower VAS scores at 6 h post-surgery in SGB-treated patients (mean [SD], 4.70 [0.91] vs 5.35 [1.32]; p = 0.040). Furthermore, patients in the SGB group experienced reduced hospital stay length (mean [SD], day, 6.61 [1.57] vs 8.72 [5.13], p = 0.042). CONCLUSIONS: Preoperative SGB emerges as a promising approach to enhance the postoperative recovery of patients undergoing laparoscopic CRC surgery. CLINICAL TRIAL REGISTRATION: ChiCTR1900028404, Principal investigator: Xia Feng, Date of registration: 12/20/2019.

A High Efficiency, Low Resistance Antibacterial Filter Formed by Dopamine-Mediated In Situ Deposition of Silver onto Glass Fibers.

The coating of filter media with silver is typically achieved by chemical deposition and aerosol processes. Whilst useful, such approaches struggle to provide uniform coating and are prone to blockage. To address these issues, an in situ method for coating glass fibers is presented via the dopamine-mediated electroless metallization method, yielding filters with low air resistance and excellent antibacterial performance. It is found that the filtration efficiency of the filters is between 94 and 97% and much higher than that of silver-coated filters produced using conventional dipping methods (85%). Additionally, measured pressure drops ranged between 100 and 150 Pa, which are lower than those associated with dipped filters (171.1 Pa). Survival rates of Escherichia coli and Bacillus subtilis bacteria exposed to the filters decreased to 0 and 15.7%±1.49, respectively after 2 h, with no bacteria surviving after 6 h. In contrast, survival rates of E. coli and B. subtilis bacteria on the uncoated filters are 92.5% and 89.5% after 6 h. Taken together, these results confirm that the in situ deposition of silver onto fiber surfaces effectively reduces pore clogging, yielding low air resistance filters that can be applied for microbial filtration and inhibition in a range of environments.

The Application of Microfibrous Entrapped Activated Carbon Composite Material for the Sarin Simulant Dimethyl Methylphosphonate Adsorption.

Granular activated carbon (GAC) has proven to be an effective adsorbent for removing the chemical warfare agent sarin (GB) and simulants like Dimethyl methylphosphonate (DMMP). However, it comes with certain limitations, including inadequate contact efficiency, notable mass transfer resistance, and lower bed utilization efficiency. This study synthesized steel fiber-entrapped activated carbon composites (SFEACs), which exhibited a maximum adsorption capacity of 285.3 mg/g at 303 K. Compared with the packed bed (PB) filled with GAC, while the adsorption capacity of SFEACS decreased, there was a substantial increase in the adsorption mass transfer rate. These SFEACs were combined with GAC to create a structural fixed bed (SFB), which demonstrated excellent performance in DMMP removal. Under identical experimental conditions, the DMMP breakthrough curve of SFB exhibited a steeper profile compared to the packed bed (PB) filled with GAC at the same bed height, and the breakthrough time against DMMP vapor could be extended by 13.8%. Furthermore, the adsorption rate constant of the Yoon-Nelson model increased by more than 17.6%, and the unused bed length, according to the Wheeler-Jonas model, decreased by more than 14%.

[Research progress on biocomposites based on bioactive glass].

Bioactive glass (BG) has been widely used in the preparation of artificial bone scaffolds due to its excellent biological properties and non-cytotoxicity, which can promote bone and soft tissue regeneration. However, due to the brittleness, poor mechanical strength, easy agglomeration and uncontrollable structure of glass material, its application in various fields is limited. In this regard, most current researches mainly focus on mixing BG with organic or inorganic materials by freeze-drying method, sol-gel method, etc., to improve its mechanical properties and brittleness, so as to increase its clinical application and expand its application field. This review introduces the combination of BG with natural organic materials, metallic materials and non-metallic materials, and demonstrates the latest technology and future prospects of BG composite materials through the development of scaffolds, injectable fillers, membranes, hydrogels and coatings. The previous studies show that the addition of BG improves the mechanical properties, biological activity and regeneration potential of the composites, and broadens the application of BG in the field of bone tissue engineering. By reviewing the recent BG researches on bone regeneration, the research potential of new materials is demonstrated, in order to provide a reference for future related research.

Review of Predicting Synergistic Drug Combinations.

The prediction of drug combinations is of great clinical significance. In many diseases, such as high blood pressure, diabetes, and stomach ulcers, the simultaneous use of two or more drugs has shown clear efficacy. It has greatly reduced the progression of drug resistance. This review presents the latest applications of methods for predicting the effects of drug combinations and the bioactivity databases commonly used in drug combination prediction. These studies have played a significant role in developing precision therapy. We first describe the concept of synergy. we study various publicly available databases for drug combination prediction tasks. Next, we introduce five algorithms applied to drug combinatorial prediction, which include traditional machine learning methods, deep learning methods, mathematical methods, systems biology methods and search algorithms. In the end, we sum up the difficulties encountered in prediction models.

A review of current effective COVID-19 testing methods and quality control.

COVID-19 is a highly infectious disease caused by the SARS-CoV-2 virus, which primarily affects the respiratory system and can lead to severe illness. The virus is extremely contagious, early and accurate diagnosis of SARS-CoV-2 is crucial to contain its spread, to provide prompt treatment, and to prevent complications. Currently, the reverse transcriptase polymerase chain reaction (RT-PCR) is considered to be the gold standard for detecting COVID-19 in its early stages. In addition, loop-mediated isothermal amplification (LMAP), clustering rule interval short palindromic repeats (CRISPR), colloidal gold immunochromatographic assay (GICA), computed tomography (CT), and electrochemical sensors are also common tests. However, these different methods vary greatly in terms of their detection efficiency, specificity, accuracy, sensitivity, cost, and throughput. Besides, most of the current detection methods are conducted in central hospitals and laboratories, which is a great challenge for remote and underdeveloped areas. Therefore, it is essential to review the advantages and disadvantages of different COVID-19 detection methods, as well as the technology that can enhance detection efficiency and improve detection quality in greater details.

Impact of nanotechnology on conventional and artificial intelligence-based biosensing strategies for the detection of viruses.

Recent years have witnessed the emergence of several viruses and other pathogens. Some of these infectious diseases have spread globally, resulting in pandemics. Although biosensors of various types have been utilized for virus detection, their limited sensitivity remains an issue. Therefore, the development of better diagnostic tools that facilitate the more efficient detection of viruses and other pathogens has become important. Nanotechnology has been recognized as a powerful tool for the detection of viruses, and it is expected to change the landscape of virus detection and analysis. Recently, nanomaterials have gained enormous attention for their value in improving biosensor performance owing to their high surface-to-volume ratio and quantum size effects. This article reviews the impact of nanotechnology on the design, development, and performance of sensors for the detection of viruses. Special attention has been paid to nanoscale materials, various types of nanobiosensors, the internet of medical things, and artificial intelligence-based viral diagnostic techniques.

Fabrication of Hierarchically Porous CuBTC@PA-PEI Composite for High-Efficiency Elimination of Cyanogen Chloride.

Cyanogen chloride (CNCl) is highly toxic and volatile, and it is difficult to effectively remove via porous substances such as activated carbon due to the weak interaction between CNCl and the adsorbent surface. Developing a highly effective elimination material against CNCl is of great importance in military chemical protection. In this work, a new metal-organic framework (MOF) CuBTC@PA-PEI (polyacrylate-polyethyleneimine) composite was prepared and exhibited excellent CNCl elimination performance in the breakthrough tests. PEI was used for the functionalization of PA with amino groups, which is beneficial to anchor with metal ions of MOF. Afterward, the growth of MOF occurred on the surface and in the pores of the matrix by molecular self-assembly via our newly proposed stepwise impregnation layer-by-layer growth method. Breakthrough tests were performed to evaluate the elimination performance of the composites against CNCl. Compared with the pristine CuBTC powder, the CuBTC@PA-PEI composite exhibited better adsorption capacity and a longer breakthrough time. By compounding with the PA matrix, a hierarchically porous structure of CuBTC@PA-PEI composite was constructed, which provides a solution to the mass transfer problem of pure microporous MOF materials. It also solves the problems of MOF molding and lays a foundation for the practical application of MOF.

Fabrication of PA-PEI-MOF303(Al) by Stepwise Impregnation Layer-by-Layer Growth for Highly Efficient Removal of Ammonia.

NH3 is a typical alkaline gaseous pollutant widely derived from industrial production and poses great risks to humans and other biota. Metal-organic frameworks (MOFs) have excellent adsorption capacities relative to materials traditionally used to adsorb NH3. However, in practice, applications of MOFs as adsorbents are restricted because of its powder form. We prepared a polyamide (PA) macroporous polyester substrate using an emulsion template method and modified the surface with polyethylenimine (PEI) to improve the MOF growth efficiency on the substrate. The difficulty of loading the MOF because of the fast nucleation rate inside the PA macroporous polyester substrate was solved using a stepwise impregnation layer-by-layer (LBL) growth method, and a PA-PEI-MOF303(Al) hierarchical pore composite that very efficiently adsorbed NH3 was successfully prepared. The PA-PEI-MOF303(Al) adsorption capacity for NH3 was 16.07 mmol·g-1 at 298 K and 100 kPa, and the PA-PEI-MOF303(Al) could be regenerated repeatedly under vacuum at 423 K. The NH3 adsorption mechanism was investigated by in situ Fourier transform infrared spectroscopy and by performing two-dimensional correlation analysis. Unlike for the MOF303(Al) powder, the formation of multi-site hydrogen bonds between Al-O-Al/C-OH, N-H, -OH, C=O, and NH3 in PA-PEI-MOF303(Al) was found to be an important reason for efficient NH3 adsorption. This study will provide a reference for the preparation of other MOF-polymer composites.

Effects of ppk1 deletion on the drug susceptibility of uropathogenic Escherichia coli producing ESBLs / 中华预防医学杂志

To investigate the effect and the mechanism of ppk1 gene deletion on the drug susceptibility of uropathogenic Escherichia coli producing extended-spectrum beta-lactamases (ESBLs-UPEC). The study was an experimental study. From March to April 2021, a strain of ESBLs-UPEC (genotype was TEM combined with CTX-M-14) named as UE210113, was isolated from urine sample of the patient with urinary tract infection in the Laboratory Department of Guangzhou Eighth People's Hospital, meanwhile its ppk1 gene knock-out strain Δpk1 and complemented strain Δpk1-C were constructed by suicide plasmid homologous recombination technique, which was used to study the effect of ppk1 gene on ESBLs-UPEC drug sensitivity and its mechanism. The drug susceptibility of UE210113, Δpk1, and Δpk1-C were measured by Vitek2 Compact System and broth microdilution method. The quantitative expression of ESBLs, outer membrane protein and multidrug efflux systems encoding genes of UE210113, Δpk1 and Δpk1-C were performed by using qRT-PCR analysis. By using two independent sample Mann-Whitney U test, the drug susceptibility results showed that, compared with UE210113 strain, the sensitivities of Δpk1 to ceftazidime, cefepime, tobramycin, minocycline and cotrimoxazole were enhanced (Z=-2.121,P<0.05;Z=-2.236,P<0.05;Z=-2.236,P<0.05;Z=-2.121,P<0.05), and the drug susceptibility of Δpk1-C restored to the same as which of UE210113 (Z=0,P>0.05). The expression levels of ESBLs-enconding genes blaTEM and blaCTX-M-14 in Δpk1 were significantly down-regulated compared with UE210113, but the expression was not restored in Δpk1-C. The expression of outer membrane protein gene omp F in Δpk1 was significantly up-regulated, while the expression of omp A and omp C were down-regulated. The results showed that the expression of multidrug efflux systems encoding genes tol C, mdt A and mdtG were down-regulated in Δpk1 compared with UE210113. The expression of all of the outer membrane protein genes and the multidrug efflux systems genes were restored in Δpk1-C. In conclusion,the lost of ppk1 gene can affect the expression of the outer membrane protein and multidrug efflux systems encoding genes of ESBLs-UPEC, which increase the sensitivity of ESBLs-UPEC to various drugs.

Effects of ppk1 deletion on the drug susceptibility of uropathogenic Escherichia coli producing ESBLs / 中华预防医学杂志

To investigate the effect and the mechanism of ppk1 gene deletion on the drug susceptibility of uropathogenic Escherichia coli producing extended-spectrum beta-lactamases (ESBLs-UPEC). The study was an experimental study. From March to April 2021, a strain of ESBLs-UPEC (genotype was TEM combined with CTX-M-14) named as UE210113, was isolated from urine sample of the patient with urinary tract infection in the Laboratory Department of Guangzhou Eighth People's Hospital, meanwhile its ppk1 gene knock-out strain Δpk1 and complemented strain Δpk1-C were constructed by suicide plasmid homologous recombination technique, which was used to study the effect of ppk1 gene on ESBLs-UPEC drug sensitivity and its mechanism. The drug susceptibility of UE210113, Δpk1, and Δpk1-C were measured by Vitek2 Compact System and broth microdilution method. The quantitative expression of ESBLs, outer membrane protein and multidrug efflux systems encoding genes of UE210113, Δpk1 and Δpk1-C were performed by using qRT-PCR analysis. By using two independent sample Mann-Whitney U test, the drug susceptibility results showed that, compared with UE210113 strain, the sensitivities of Δpk1 to ceftazidime, cefepime, tobramycin, minocycline and cotrimoxazole were enhanced (Z=-2.121,P<0.05;Z=-2.236,P<0.05;Z=-2.236,P<0.05;Z=-2.121,P<0.05), and the drug susceptibility of Δpk1-C restored to the same as which of UE210113 (Z=0,P>0.05). The expression levels of ESBLs-enconding genes blaTEM and blaCTX-M-14 in Δpk1 were significantly down-regulated compared with UE210113, but the expression was not restored in Δpk1-C. The expression of outer membrane protein gene omp F in Δpk1 was significantly up-regulated, while the expression of omp A and omp C were down-regulated. The results showed that the expression of multidrug efflux systems encoding genes tol C, mdt A and mdtG were down-regulated in Δpk1 compared with UE210113. The expression of all of the outer membrane protein genes and the multidrug efflux systems genes were restored in Δpk1-C. In conclusion,the lost of ppk1 gene can affect the expression of the outer membrane protein and multidrug efflux systems encoding genes of ESBLs-UPEC, which increase the sensitivity of ESBLs-UPEC to various drugs.

A Review on Cyanide Gas Elimination Methods and Materials.

Cyanide gas is highly toxic and volatile and is among the most typical toxic and harmful pollutants to human health and the environment found in industrial waste gas. In the military context, cyanide gas has been used as a systemic toxic agent. In this paper, we review cyanide gas elimination methods, focusing on adsorption and catalysis approaches. The research progress on materials capable of affecting cyanide gas adsorption and catalytic degradation is discussed in depth, and the advantages and disadvantages of various materials are summarized. Finally, suggestions are provided for future research directions with respect to cyanide gas elimination materials.

ReDisX, a machine learning approach, rationalizes rheumatoid arthritis and coronary artery disease patients uniquely upon identifying subpopulation differentiation markers from their genomic data.

Diseases originate at the molecular-genetic layer, manifest through altered biochemical homeostasis, and develop symptoms later. Hence, symptomatic diagnosis is inadequate to explain the underlying molecular-genetic abnormality and individual genomic disparities. The current trends include molecular-genetic information relying on algorithms to recognize the disease subtypes through gene expressions. Despite their disposition toward disease-specific heterogeneity and cross-disease homogeneity, a gap still exists in describing the extent of homogeneity within the heterogeneous subpopulation of different diseases. They are limited to obtaining the holistic sense of the whole genome-based diagnosis resulting in inaccurate diagnosis and subsequent management. Addressing those ambiguities, our proposed framework, ReDisX, introduces a unique classification system for the patients based on their genomic signatures. In this study, it is a scalable machine learning algorithm deployed to re-categorize the patients with rheumatoid arthritis and coronary artery disease. It reveals heterogeneous subpopulations within a disease and homogenous subpopulations across different diseases. Besides, it identifies granzyme B (GZMB) as a subpopulation-differentiation marker that plausibly serves as a prominent indicator for GZMB-targeted drug repurposing. The ReDisX framework offers a novel strategy to redefine disease diagnosis through characterizing personalized genomic signatures. It may rejuvenate the landscape of precision and personalized diagnosis and a clue to drug repurposing.

The Relationship between Osteoinduction and Vascularization: Comparing the Ectopic Bone Formation of Five Different Calcium Phosphate Biomaterials.

Objective: The objective of this study is to compare the bone induction of five kinds of calcium phosphate (Ca-P) biomaterials implanted in mice and explore the vascularization and particle-size-related osteoinductive mechanism. Methods: The following five kinds of Ca-P biomaterials including hydroxyapatite (HA) and/or tricalcium phosphate (TCP) were implanted in the muscle of 30 BALB/c mice (n = 6): 20 nm HA (20HA), 60 nm HA (60HA), 12 µm HA (12HA), 100 nm TCP (100TCP) and 12 µm HA + 100 nm TCP (HATCP). Then, all animals were put on a treadmill to run 30 min at a 6 m/h speed each day. Five and ten weeks later, three mice of each group were killed, and the samples were harvested to assess the osteoinductive effects by hematoxylin eosin (HE), Masson's trichrome and safranine−fast green stainings, and the immunohistochemistry of the angiogenesis and osteogenesis markers CD31 and type I collagen (ColI). Results: The numbers of blood vessels were 139 ± 29, 118 ± 25, 78 ± 15, 65 ± 14 in groups HATCP, 100TCP, 60HA and 20HA, respectively, which were significantly higher than that of group 12HA (12 ± 5) in week 5 (p < 0.05). The area percentages of new bone tissue were (7.33 ± 1.26)% and (8.49 ± 1.38)% in groups 100TCP and HATCP, respectively, which were significantly higher than those in groups 20HA (3.27 ± 0.38)% and 60HA (3.43 ± 0.27)% (p < 0.05); however, no bone tissue was found in group 12HA 10 weeks after transplantation. The expression of CD31 was positive in new blood vessels, and the expression of ColI was positive in new bone tissue. Conclusions: Nanoscale Ca-P biomaterials could induce osteogenesis in mice muscle, and the osteoinductive effects of TCP were about 124% higher than those of 20HA and 114% higher than those of 60HA. The particle size of the biomaterials affected angiogenesis and osteogenesis. There was a positive correlation between the number of blood vessels and the area percentage of new bone tissue; therefore, osteoinduction is closely related to vascularization. Our results provide an experimental basis for the synthesis of calcium−phosphorus matrix composites and for further exploration of the osteoinductive mechanism.

Comparison of the effect of bone induction with different exercise modes in mice.

BACKGROUD: Calcium phosphate biomaterials have excellent bone inductivity, and exercise can promote the bone formation of biomaterials in animals, but it is not clear which exercise mode is better. OBJECTIVE: To explore the effect of different exercise modes on osteoinduction by calcium phosphate-based biomaterials which were implanted in mice. METHOD: The collagen-thermosensitive hydrogel-calcium phosphate (CTC) composite was prepared and transplanted in the thigh muscle of mice, then all mice were divided randomly into four groups (n = 10): the uphill running group, the downhill running group, the swimming group and the control group (conventional breeding). Ten weeks later, the samples were harvested, fixed, decalcified, embedded in paraffin and stained with hematoxylin and eosin (H&E), and then the osteoinduction phenomenon was observed and compared through digital slice scanning system. The area percentage of new bone-related tissues and the number of osteocytes and chondrocytes were counted and calculated. Lastly, the immunohistochemistry of type I collagen (ColI) and osteopontin (OPN) was performed to identify the new bone tissues. RESULTS: The area percentage of new bone-related tissues and the number of osteocytes and chondrocytes were positively correlated; ordering from most to least of each group were as followings: the uphill running group > the swimming group > the downhill running group > the control group. The immunostaining of ColI and OPN results showed that both of the two proteins were identified in the new bone tissues, indicating that the CTC composite could induce ectopic bone formation in mice, especially training for uphill running and swimming. CONCLUSION: Our results show that uphill running or swimming is a form of exercise that is beneficial to osteogenesis. According to this, we propose treatment with artificial bone transplantation to patients who suffer from bone defects. Patients should do moderate exercise, such as running uphill on the treadmill or swimming.

Artificial intelligence: A powerful paradigm for scientific research.

Artificial intelligence (AI) coupled with promising machine learning (ML) techniques well known from computer science is broadly affecting many aspects of various fields including science and technology, industry, and even our day-to-day life. The ML techniques have been developed to analyze high-throughput data with a view to obtaining useful insights, categorizing, predicting, and making evidence-based decisions in novel ways, which will promote the growth of novel applications and fuel the sustainable booming of AI. This paper undertakes a comprehensive survey on the development and application of AI in different aspects of fundamental sciences, including information science, mathematics, medical science, materials science, geoscience, life science, physics, and chemistry. The challenges that each discipline of science meets, and the potentials of AI techniques to handle these challenges, are discussed in detail. Moreover, we shed light on new research trends entailing the integration of AI into each scientific discipline. The aim of this paper is to provide a broad research guideline on fundamental sciences with potential infusion of AI, to help motivate researchers to deeply understand the state-of-the-art applications of AI-based fundamental sciences, and thereby to help promote the continuous development of these fundamental sciences.

DeepDRIM: a deep neural network to reconstruct cell-type-specific gene regulatory network using single-cell RNA-seq data.

Single-cell RNA sequencing has enabled to capture the gene activities at single-cell resolution, thus allowing reconstruction of cell-type-specific gene regulatory networks (GRNs). The available algorithms for reconstructing GRNs are commonly designed for bulk RNA-seq data, and few of them are applicable to analyze scRNA-seq data by dealing with the dropout events and cellular heterogeneity. In this paper, we represent the joint gene expression distribution of a gene pair as an image and propose a novel supervised deep neural network called DeepDRIM which utilizes the image of the target TF-gene pair and the ones of the potential neighbors to reconstruct GRN from scRNA-seq data. Due to the consideration of TF-gene pair's neighborhood context, DeepDRIM can effectively eliminate the false positives caused by transitive gene-gene interactions. We compared DeepDRIM with nine GRN reconstruction algorithms designed for either bulk or single-cell RNA-seq data. It achieves evidently better performance for the scRNA-seq data collected from eight cell lines. The simulated data show that DeepDRIM is robust to the dropout rate, the cell number and the size of the training data. We further applied DeepDRIM to the scRNA-seq gene expression of B cells from the bronchoalveolar lavage fluid of the patients with mild and severe coronavirus disease 2019. We focused on the cell-type-specific GRN alteration and observed targets of TFs that were differentially expressed between the two statuses to be enriched in lysosome, apoptosis, response to decreased oxygen level and microtubule, which had been proved to be associated with coronavirus infection.

Dissociation of water on atomic oxygen-covered Rh nanoclusters supported on graphene/Ru(0001).

We studied the dissociation of water (H2O*, with * denoting adspecies) on atomic oxygen (O*)-covered Rh nanoclusters (RhO* ) supported on a graphene film grown on a Ru(0001) surface [G/Ru(0001)] under ultrahigh-vacuum conditions and with varied surface-probe techniques and calculations based on density-functional theory. The graphene had a single rotational domain; its lattice expanded by about 5.7% to match the Ru substrate structurally better. The Rh clusters were grown by depositing Rh vapors onto G/Ru(0001); they had an fcc phase and grew in (111) orientation. Water adsorbed on the Rh clusters was dissociated exclusively in the presence of O*, like that on a Rh(111) single-crystal surface. Contrary to the case on Rh(111)O* , excess O* (even at a saturation level) on small RhO* clusters (diameter of 30-34 Å) continued to promote, instead of inhibiting, the dissociation of water; the produced hydroxyl (OH*) increased generally with the concentration of O* on the clusters. The difference results from more reactive O* on the RhO* clusters. O* on RhO* clusters activated the dissociation via both the formation of hydrogen bonds with H2O* and abstraction of H directly from H2O*, whereas O* on Rh(111)O* assisted the dissociation largely via the formation of hydrogen bonds, which was readily obstructed with an increased O* coverage. As the disproportionation (2 OH* → H2O* + O*) is endothermic on the RhO* clusters but exothermic on Rh(111)O* , OH* produced on RhO* clusters showed a thermal stability superior to that on the Rh(111)O* surface-thermally stable up to 400 K.