Search

Heterogeneous deep learning clusters commonly host a variety of distributed learning jobs. In such scenarios, the training efficiency of learning models is negatively affected by the slowest worker. To accelerate the training process, multiple learning jobs may compete for limited computational resources, posing significant challenges to multi-job placement among heterogeneous workers. This article presents a heterogeneity-aware scheduler to solve the multi-job placement problem while taking into account job sizing and load balancing, minimizing the average Job Completion Time (JCT) of deep learning jobs. A novel scheme based on proportional training workload assignment, feasible solution categorization, and matching markets is proposed with theoretical guarantees. To further reduce the computational complexity for low latency decision-making and improve scheduling fairness, we propose to construct the sparsification of feasible solution categories through sampling, which has negligible performance loss in JCT. We evaluate the performance of our design with real-world deep neural network benchmarks on heterogeneous computing clusters. Experimental results show that, compared to existing solutions, the proposed sampling-based scheme can achieve 1) results within 2.04% of the optimal JCT with orders-of-magnitude improvements in algorithm running time, and 2) high scheduling fairness among learning jobs.

The monocular visual-inertial odometry (VIO) based on the direct method can leverage all available pixels in the image to simultaneously estimate the camera motion and reconstruct the denser map of the scene in real-time. However, the direct method is sensitive to photometric changes, which can be compensated by introducing geometric information in the environment. In this article, we propose a monocular direct sparse VIO, which exploits the planar regularities (PVI-DSO). Our system detects the planar regularities from the 3-D mesh built on the estimated map points. To improve the pose estimation accuracy with the geometric information, a tightly coupled coplanar constraint expression is used to express photometric error in the direct method. In addition, to improve the optimization efficiency, we elaborately derive the analytical Jacobian of the linearization form for the coplanar constraint. Finally, the inertial measurement error, coplanar point photometric error, noncoplanar photometric error, and prior error are added into the optimizer, which simultaneously improves the pose estimation accuracy and mesh itself. We verified the performance of the whole system on the simulation and real-world datasets. Extensive experiments have demonstrated that our system outperforms the state-of-the-art counterparts.

Systemic lupus erythematosus (SLE) is a debilitating autoimmune disorder characterized by unknown pathogenesis and heterogeneous clinical manifestations. The current existing serum biomarkers for SLE have limited sensitivity or specificity, making early and precise diagnosis difficult. Here, we identified two N-glycans on serum immunoglobulin G (IgG) as excellent diagnostic biomarkers for SLE based on in-depth glycomic analyses of 389 SLE patients and 304 healthy controls. These two N-glycan biomarkers are specific for diagnosing SLE, as no significant changes in these biomarkers were observed in other systemic autoimmune diseases that are easily confused with SLE, such as rheumatoid arthritis, primary Sjögren’s syndrome, or systemic sclerosis. Notably, the two N-glycan biomarkers proved to be autoantibody-independent and all-stage patient suitable. The two N-glycan biomarkers are demonstrated to be located on the Fc region based on fragment-specific glycan analysis and glycopeptide analysis, suggesting their close correlation with disease activity. Enzyme analyses revealed dysregulation of a series of glycotransferases in SLE, which might be responsible for the observed glycan alteration. Our findings provide insights into efficient population screening based on serum IgG glycosylation and potential new pathogenic factors of SLE.

ABSTRACTSevere hydronephrosis increases the risk of urinary tract infection and irretrievable renal fibrosis. While TGFβ1-mediated fibrotic changes in proximal tubular epithelial cells and fatty acid oxidation (FAO) deregulation contribute to renal fibrosis and hydronephrosis. Firstly, a few elements were analyzed in this paper, including differentially-expressed long non-coding RNAs (lncRNAs), and miRNAs correlated to CPT1A, RXRA, and NCOA1. This paper investigated TGFβ1 effects on lncRNA FABP5P3, CPT1A, RXRA, and NCOA1 expression and fibrotic changes in HK-2 cells and FABP5P3 overexpression effects on TGFβ1-induced changes. Moreover, this paper predicted and proved that miR-22 binding to lncRNA FABP5P3, 3’UTR of CPT1A, RXRA, and NCOA1 was validated. The dynamic effects of the FABP5P3/miR-22 axis on TGFβ1-induced changes were investigated. A Renal fibrosis model was established in unilateral ureteral obstruction (UUO) mice, and FABP5P3 effects were investigated. Eventually, this paper concluded that TGFβ1 inhibited lncRNA FABP5P3, CPT1A, RXRA, and NCOA1 expression, induced fibrotic changes in HK-2 cells, and induced metabolic reprogramming within HK-2 cells, especially lower FAO. FABP5P3 overexpression partially reversed TGFβ1-induced changes. miR-22 targeted lncRNA FABP5P3, CPT1A, RXRA, and NCOA1. LncRNA FABP5P3 counteracted miR-22 inhibition of CPT1A, NCOA1, and RXRA through competitive binding. TGFβ1 stimulation induced the activation of TGFβ/SMAD and JAG/Notch signaling pathways; Nocth2 knockdown reversed TGFβ1 suppression on lncRNA FABP5P3. FABP5P3 overexpression attenuated renal fibrosis in unilateral ureteral obstruction mice. The LncRNA FABP5P3/miR-22 axis might be a potent target for improving the FAO deregulation and fibrotic changes in proximal TECs under TGFβ1 stimulation.

Postoperative abdominal adhesion is a very common and serious complication, resulting in pain, intestinal obstruction and heavy economic burden. Post-injury inflammation that could activate the coagulation cascade and deposition of fibrin is a major cause of adhesion. Many physical barrier membranes are used to prevent abdominal adhesion, but their efficiency is limited due to the lack of anti-inflammatory activity. Here, an electrospinning membrane composed of poly(lactic-co-glycolic acid) (PLGA) providing support and mechanical strength and chondroitin sulfate (CS) conferring anti-inflammation activity is fabricated for preventing abdominal adhesion after injury. The PLGA/CS membrane shows a highly dense fiber network structure with improved hydrophilicity and good cytocompatibility. Importantly, the PLGA/CS membrane with a mass ratio of CS at 20% provides superior anti-adhesion efficiency over a native PLGA membrane and commercial poly(d, l-lactide) (PDLLA) film in abdominal adhesion trauma rat models. The mechanism is that the PLGA/CS membrane could alleviate the local inflammatory response as indicated by the promoted percentage of anti-inflammatory M2-type macrophages and decreased expression of pro-inflammatory factors, such as IL-1β, TNF-α and IL-6, resulting in the suppression of the coagulation system and the activation of the fibrinolytic system. Furthermore, the deposition of fibrin at the abdominal wall was inhibited, and the damaged abdominal tissue was repaired with the treatment of the PLGA/CS membrane. Collectively, the PLGA/CS electrospinning membrane is a promising drug-/cytokine-free anti-inflammatory barrier for post-surgery abdominal adhesion prevention and a bioactive composite for tissue regeneration.

The role of glucocorticoids in oogenesis remains to be elucidated. cyp11c1encodes the key enzyme involved in the synthesis of cortisol, the major glucocorticoid in teleosts. In our previous study, we mutated cyp11c1in tilapia and analyzed its role in spermatogenesis. In this study, we analyzed its role in oogenesis. cyp11c1+/−XX tilapia showed normal ovarian morphology but poor egg quality, as indicated by the mortality of embryos before 3 d post fertilization, which could be partially rescued by the supplement of exogenous cortisol to the mother fish. Transcriptome analyses revealed reduced expression of maternal genes in the eggs of the cyp11c1+/−XX fish. The cyp11c1−/−females showed impaired vitellogenesis and arrested oogenesis due to significantly decreased serum cortisol. Further analyses revealed decreased serum E2 level and expression of amh, an important regulator of follicular cell development, and increased follicular cell apoptosis in the ovaries of cyp11c1−/−XX fish, which could be rescued by supplement of either exogenous cortisol or E2. Luciferase assays revealed a direct regulation of cortisol and E2 on amhtranscription via GRs or ESRs. Taken together, our results demonstrate that cortisol safeguards oogenesis by promoting follicular cell survival probably via Amh signaling.

Cold stress significantly limits the yield and quality of tomato. Deciphering the key genes related to cold tolerance is important for selecting and breeding superior cold-tolerant varieties. γ-aminobutyric acid (GABA) responds to various types of stress by rapidly accumulating in plant. In this study, glutamic acid decarboxylase (GAD2) was a positive regulator to enhance cold stress tolerance of tomato. Overexpression of SlGAD2decreased the extent of cytoplasmic membrane damage and increased the endogenous GABA content, antioxidant enzyme activities, and reactive oxygen species (ROS) scavenging capacity in response to cold stress, whereas Slgad2mutant plants showed the opposite trend. In addition, SlGAD2induced anthocyanin biosynthesis in response to cold stress by increasing the content of endogenous GABA. Further study revealed that SlGAD2expression was negatively regulated by the transcription factor SlTHM27. However, the transcript levels of SlTHM27were repressed under cold stress. Antioxidant enzyme activities, SlGAD2transcript levels, GABA and anthocyanin contents were significantly increased in Slthm27mutant plants. Further, our study demonstrated that SlTHM27 decreases SlGAD2-promoted cold resistance in tomato by repressing SlGAD2transcription. Overall, our results showed that the SlTHM27-SlGAD2model regulates the cold tolerance in tomato by regulating GABA and anthocyanin.

The impact of sex chromosomes and their turnover in speciation remains a subject of ongoing debate in the field of evolutionary biology. Fishes are the largest group of vertebrates, and they exhibit unparalleled sexual plasticity, as well as diverse sex-determining (SD) genes, sex chromosomes, and sex-determination mechanisms. This diversity is hypothesized to be associated with the frequent turnover of sex chromosomes in fishes. Although it is evident that amhand amhr2are repeatedly and independently recruited as SD genes, their relationship with the rapid turnover of sex chromosomes and the biodiversity of fishes remains unknown. We summarize the canonical models of sex chromosome turnover and highlight the vital roles of gene mutation and hybridization with empirical evidence. We revisit Haldane's rule and the large X-effect and propose the hypothesis that sex chromosomes accelerate speciation by multiplying genotypes via hybridization. By integrating recent findings on the turnover of SD genes, sex chromosomes, and sex-determination systems in fish species, this review provides insights into the relationship between sex chromosome evolution and biodiversity in fishes.

Lithium-ion battery state-of-charge (SOC) estimation plays an indispensable role in the battery management system functionalities. To achieve SOC estimation with high accuracy, a precise battery model is required. Among various battery models, electrochemical model (EM) stands out for its ability to provide insights into the electrochemical mechanisms of the battery. However, the model parameters will vary with the battery running and aging, besides, the model-based methods cannot well handle the model errors. Therefore, to ensure estimation accuracy, selecting a more insightful battery model and a more effective estimation algorithm becomes crucial. This paper proposes an electrochemical model-based SOC estimation method with online parameter adjustment and an optimized algorithm. Firstly, a single particle model (SPM) with electrolyte and thermal dynamics is introduced and simplified by an approximate solution. Secondly, the particle swarm optimization (PSO) method is used to solve the problem of the particle filter (PF) algorithm, such as particle degradation and particle diversity decrease. Finally, during SOC estimating, a double-scale dual particle filter (D-PF) is utilized to adjust parameters as battery aging and temperature changes. The battery tests are applied under different working conditions and temperatures. Experimental results show that the error of SOC estimation reduces significantly compared to the previous electrochemical model-based method, and the proposed method has a higher accuracy and effectiveness.

Climatic extreme events will occur more frequently, which will seriously threaten the life safety of residents in arid and semi-arid areas. Cairo, Algiers and Dubai are located at the crossroads of the global climate pattern. Climatic extreme events will seriously threaten the lives of local residents. Therefore, it is of great value to study the impact of the Marine Cloud Brightening (MCB) scheme on these metropolises. Based on the HadGEM2-ES model simulations of the MCB scheme, we used six climatic extreme indices, the hottest days (TXx), the coolest nights (TNn), the warm spell duration (WSDI), the cold spell duration (CSDI), the consecutive dry days (CDD) and wettest consecutive five days (RX5day), to analyze the impact of MCB schemes on the extreme climate in Cairo, Algiers and Dubai in this study. Compared with an RCP4.5 scenario, the implementation of the MCB scheme from 2030 to 2059 would increase the mean annual RX5day index in Cairo by 1.98mm, decrease the average annual WSDI index in Algiers by 26 days and decrease the average annual WSDI index in Dubai by 70 days. The implementation of the MCB scheme is of great significance to mitigate climatic extreme events in Cairo, Algiers and Dubai.

Introduction Cervical squamous cell carcinoma (CESC) is the most common gynecological malignancy worldwide. Although the cancer susceptibility 18 (CASC18)gene was involved in the regulation of cancer biology, its specific role in CESC is not well characterized.Methods CASC18-related axis was predicted by bioinformatic analyses, and the competing endogenous RNA (ceRNA) interaction was further validated using quantitative real-time PCR, western blotting, RNA pulldown, and luciferase reporter assays. Transwell and wound healing assays were performed to verify the effect of CASC18on SiHa and HeLa cell motility.Results We found that CASC18was upregulated in CESC tissues. Moreover, interference with CASC18attenuated NUAK1-mediated epithelial-mesenchymal transition (EMT) and thus suppressed cancer cell motility. Furthermore, the effects of CASC18knockdown on CESC cells were partly rescued by transfection with the miR-5586-5p inhibitor. Additionally, our findings indicated that CASC18acts as a ceRNA to enhance NUAK1expression by sponging miR-5586-5p.Conclusion Our study showed a novel CASC18/miR-5586-5p/NUAK1ceRNA axis that could regulate cell invasion and migration by modulating EMT in CESC. These findings suggest that CASC18may potentially serve as a novel therapeutic target in CESC treatment.

Time delay-based the 5th Generation Mobile Communication Technology (5G) positioning is a main method to perform high-precision positioning in Global Navigation Satellite System (GNSS) denied areas. However, in practical applications, the occlusion of signals in a complex environment results in few observable base stations, which affects the reliability and accuracy of positioning. The aim of this study is to improve the performance of the 5G positioning in complex environments with an insufficient number of observable base stations. First, the Angle of Departure (AOD) capability of multi-antennas is integrated into Multi-Round-Trip-Time (Multi-RTT) positioning, establishing a novel 5G RTT/AOD positioning model. Then, the influencing factors of positioning performance, including the Dilution of Precision (DOP) and the accuracy of the AOD measurements, is analyzed. The relationship between DOP and RTT/AOD positioning accuracy is deduced. Afterwards, simulation experiments are performed on 5G positioning with the Multi-RTT and RTT/AOD methods in two scenarios with good and complex environments. The theoretical analysis and experimental results show that 5G positioning with the RTT/AOD method increases the horizontal and vertical accuracies by approximately 25 and 65%, respectively, compared with the Multi-RTT method. The positioning reliability is also greatly improved. The proposed model can well solve the inefficiency of 5G positioning with the RTT method in scenarios where the number of base stations is less than three.

High-impact polystyrene (HIPS)/organophilic alpha-zirconium phosphate (organophilic α-ZrP, OZrP) nanocomposites were prepared by melt blending method. The XRD and high-resolution transmission electronic microscopy results showed that an intercalated/exfoliated structure of nanocomposites was formed. The thermal degradation behaviors indicated that the HIPS/OZrP nanocomposites displayed similar thermo-oxidative stability with that of HIPS, and α-ZrP could promote the formation of char residue of the nanocomposites. The combustion behaviors revealed that the addition of α-ZrP could reduce the value of peak heat release rate (HRR) and mass loss rate (MLR) of the nanocomposites. The peak HRR value of HIPS/OZrP nanocomposites with 3 wt% OZrP was 17 % lower than that of pure HIPS. The trends of MLR were the same as those of the HRR, which indicated that the mechanism of the observed reduction both in HRR and MLR by addition of the layered phosphate depended mainly on the condensed phase flame-retardant process instead of the gas phase flame-retardant process. It could be observed that HIPS left no residues after burning, whereas the HIPS/OZrP nanocomposites left the consistent and compact char residues. Moreover, graphite sheets and carbon nanotubes were observed in the char residues of nanocomposites. The experimental results suggested that polymers would have improved fire performance with the high crystallization of char residue in the presence of α-ZrP.

Terahertz time-domain spectroscopy (THz-TDS) as a new non-destructive testing method has been applied to the identification of 41 official and unofficial rhubarb samples in the present work. The THz time domain spectra of rhubarb samples were preprocessed and then used to establish an identification model by using fuzzy rule-building expert systems (FuRES). The model was validated using a bootstrapped Latin-partitions (BLPs) method with 10 bootstraps and 4 Latin-partitions. The obtained results showed that the model has good predictive ability with respect to the classification accuracy of 94.8 ± 0.5% and 95.2 ± 0.1% by using the preprocessing methods of Savitzky–Golay (S–G) first derivative combined with either one of two orthogonal signal correction (OSC) methods, respectively. The proposed method showed that THz-TDS combined with chemometrics can be used to identify genuine and counterfeit Chinese herbal medicines, as well as official and unofficial rhubarb.

The development of 20(S)-ginsenoside Rh2 (Rh2s) as a chemoprevention agent is limited by its low oral bioavailability. The goals of this study were to determine the mechanisms responsible for its poor oral absorption and to improve its bioavailability by overcoming the barrier to its absorption. Comprehensive studies were conducted using the following models: 1) monolayers of Caco-2, parental, and multidrug resistance gene (MDR1)-overexpressing Madin-Darby canine kidney II (MDCKII) cells; 2) pharmacokinetics in wild-type (WT) FVB, MDR1a/b knockout [MDR1a/b(–/–)] FVB, and A/J mice; and 3) intestinal perfusion in WT, MDR1a/b(–/–) FVB, and A/J mice. Two P-glycoprotein (P-gp) inhibitors, verapamil and cyclosporine A, substantially decreased the efflux ratio of Rh2s from 28.5 to 1.0 and 1.2, respectively, in Caco-2 cells. The intracellular concentrations of Rh2s were also significantly increased (2.3- and 3.9-fold) in the presence of inhibitors. Similar results were obtained when transcellular transport of Rh2s were determined using MDR1-overexpressing MDCKII cells in the absence or presence of cyclosporine A. Compared with WT mice, the plasma Cmaxand AUC0–∞of Rh2s were substantially increased by 17- and 23-fold in MDR1a/b(–/–) FVB mice, respectively. In the A/J mice, the oral bioavailability of Rh2s (0.94% at 5 mg/kg and 0.52% at 20 mg/kg) was substantially increased by P-gp inhibitor to 33.18 and 27.14%, respectively. As expected, deletion or inhibition of P-gp significantly increased absorption and steady-state plasma concentration of Rh2s in a mouse intestinal perfusion model. In conclusion, Rh2s is a good substrate of P-gp, and inhibition of P-gp can significantly enhance its oral bioavailability.

We characterized the in vitro glucuronidation of prunetin, a prodrug of genistein that is a highly active cancer prevention agent. Metabolism studies were conducted using expressed human UGT isoforms and microsomes/S9 fractions prepared from intestine and liver of rodents and humans. The results indicated that human intestinal microsomes were more efficient than liver microsomes in glucuronidating prunetin, but rates of metabolism were dependent on time of incubation at 37 °C. Human liver and intestinal microsomes mainly produced metabolite 1 (prunetin-5- O-glucuronide) and metabolite 2 (prunetin-4′- O-glucuronide), respectively. Using 12 human UGT isoforms, we showed that UGT1A7, UGT1A8, and UGT1A9 were mainly responsible for the formation of metabolite 1, whereas UGT1A1, UGT1A8, and UGT1A10 were mainly responsible for the formation of metabolite 2. This isoform-specific metabolism was consistent with earlier results obtained using human liver and intestinal microsomes, as the former (liver) is UGT1A9-rich whereas the latter is UGT1A10-rich. Surprisingly, we found that the thermostability of the microsomes was isoform- and organ-dependent. For example, human liver microsomal UGT activities were much more heat-stable (37 °C) than intestinal microsomal UGT activities, consistent with the finding that human UGT1A9 is much more thermostable than human UGT1A10 and UGT1A8. The organ-specific thermostability profiles were also evident in rat microsomes and mouse S9 fractions, even though human intestinal glucuronidation of prunetin differs significantly from rodent intestinal glucuronidation. In conclusion, prunetin glucuronidation is species-, organ-, and UGT-isoform-dependent, all of which may be impacted by the thermostability of specific UGT isoforms involved in the metabolism.

HT-7U is a fully superconducting tokamak. All of the toroidal field and poloidal field superconducting magnets are made of NbTi superconductor and are designed to operate at 3.8 K in the steady-state operation mode. The magnet system will be cooled with forced-flow supercritical helium to this temperature level with an equivalent refrigeration capacity of ~2 kW/4.4 K. To satisfy this requirement, a helium refrigerator is designed to be able to provide the cooling power at 3.5, 4.5, and 80 K for the coils, supports, and thermal shields. The refrigerator can also produce liquid helium for the cooling of the current leads. This paper describes the cryogenic system of HT-7U, the refrigeration process and helium plant, and the forced-flow cooling requirement of the superconducting magnets.