Beyond Conventional Charge Density Wave for Strongly Enhanced 2D Superconductivity in 1H-TaS2 Superlattices.

Noncentrosymmetric transition metal dichalcogenide (TMD) monolayers offer a fertile platform for exploring unconventional Ising superconductivity (SC) and charge density waves (CDWs). However, the vulnerability of isolated monolayers to structural disorder and environmental oxidation often degrade their electronic coherence. Herein, an alternative approach is reported for fabricating stable and intrinsic monolayers of 1H-TaS2 sandwiched between SnS blocks in a (SnS)1.15TaS2 van der Waals (vdW) superlattice. The SnS block layers not only decouple individual 1H-TaS2 sublayers to endow them with monolayer-like electronic characteristics, but also protect the 1H-TaS2 layers from electronic degradation. The results reveal the characteristic 3 × 3 CDW order in 1H-TaS2 sublayers associated with electronic rearrangement in the low-lying sulfur p band, which uncovers a previously undiscovered CDW mechanism rather than the conventional Fermi surface-related framework. Additionally, the (SnS)1.15TaS2 superlattice exhibits a strongly enhanced Ising-like SC with a layer-independent Tc of ≈3.0 K, comparable to that of the isolated monolayer 1H-TaS2 sample, presumably attributed to their monolayer-like characteristics and retained Fermi states. These results provide new insights into the long-debated CDW order and enhanced SC of monolayer 1H-TaS2, establishing bulk vdW superlattices as promising platforms for investigating exotic collective quantum phases in the 2D limit.

A Fast Self-Calibration Method for Dual-Axis Rotational Inertial Navigation Systems Based on Invariant Errors.

In order to ensure that dual-axis rotational inertial navigation systems (RINSs) maintain a high level of accuracy over the long term, there is a demand for periodic calibration during their service life. Traditional calibration methods for inertial measurement units (IMUs) involve removing the IMU from the equipment, which is a laborious and time-consuming process. Reinstalling the IMU after calibration may introduce new installation errors. This paper focuses on dual-axis rotational inertial navigation systems and presents a system-level self-calibration method based on invariant errors, enabling high-precision automated calibration without the need for equipment disassembly. First, navigation parameter errors in the inertial frame are expressed as invariant errors. This allows the corresponding error models to estimate initial attitude even more rapidly and accurately in cases of extreme misalignment, eliminating the need for coarse alignment. Next, by utilizing the output of a gimbal mechanism, angular velocity constraint equations are established, and the backtracking navigation is introduced to reuse sensor data, thereby reducing the calibration time. Finally, a rotation scheme for the IMU is designed to ensure that all errors are observable. The observability of the system is analyzed based on a piecewise constant system method and singular value decomposition (SVD) observability analysis. The simulation and experimental results demonstrate that this method can effectively estimate IMU errors and installation errors related to the rotation axis within 12 min, and the estimated error is less than 4%. After using this method to compensate for the calibration error, the velocity and position accuracies of a RINS are significantly improved.

Highly Selective On-Surface Ring-Opening of Aromatic Azulene Moiety.

Controllable ring-opening of polycyclic aromatic hydrocarbons plays a crucial role in various chemical and biological processes. However, breaking down aromatic covalent C-C bonds is exceptionally challenging due to their high stability and strong aromaticity. This study presents a seminal report on the precise and highly selective on-surface ring-opening of the seven-membered ring within the aromatic azulene moieties under mild conditions. The chemical structures of the resulting products were identified using bond-resolved scanning probe microscopy. Furthermore, through density functional theory calculations, we uncovered the mechanism behind the ring-opening process and elucidated its chemical driving force. The key to achieving this ring-opening process lies in manipulating the local aromaticity of the aromatic azulene moiety through strain-induced internal ring rearrangement and cyclodehydrogenation. By precisely controlling these factors, we successfully triggered the desired ring-opening reaction. Our findings not only provide valuable insights into the ring-opening process of polycyclic aromatic hydrocarbons but also open up new possibilities for the manipulation and reconstruction of these important chemical structures.

Molecular regulation of retinal regeneration is context specific.

Many genes are known to regulate retinal regeneration following widespread tissue damage. Conversely, genes controlling regeneration following limited retinal cell loss, akin to disease conditions, are undefined. Combining a novel retinal ganglion cell (RGC) ablation-based glaucoma model, single cell omics, and rapid CRISPR/Cas9-based knockout methods to screen 100 genes, we identified 18 effectors of RGC regeneration kinetics. Surprisingly, 32 of 33 previously known/implicated regulators of retinal tissue regeneration were not required for RGC replacement; 7 knockouts accelerated regeneration, including sox2, olig2, and ascl1a . Mechanistic analyses revealed loss of ascl1a increased "fate bias", the propensity of progenitors to produce RGCs. These data demonstrate plasticity and context-specificity in how genes function to control regeneration, insights that could help to advance disease-tailored therapeutics for replacing lost retinal cells. One sentence summary: We discovered eighteen genes that regulate the regeneration of retinal ganglion cells in zebrafish.

Common and divergent gene regulatory networks control injury-induced and developmental neurogenesis in zebrafish retina.

Following acute retinal damage, zebrafish possess the ability to regenerate all neuronal subtypes through Müller glia (MG) reprogramming and asymmetric cell division that produces a multipotent Müller glia-derived neuronal progenitor cell (MGPC). This raises three key questions. First, do MG reprogram to a developmental retinal progenitor cell (RPC) state? Second, to what extent does regeneration recapitulate retinal development? And finally, does loss of different retinal cell subtypes induce unique MG regeneration responses? We examined these questions by performing single-nuclear and single-cell RNA-Seq and ATAC-Seq in both developing and regenerating retinas. Here we show that injury induces MG to reprogram to a state similar to late-stage RPCs. However, there are major transcriptional differences between MGPCs and RPCs, as well as major transcriptional differences between activated MG and MGPCs when different retinal cell subtypes are damaged. Validation of candidate genes confirmed that loss of different subtypes induces differences in transcription factor gene expression and regeneration outcomes.

Common and divergent gene regulatory networks control injury-induced and developmental neurogenesis in zebrafish retina.

Following acute retinal damage, zebrafish possess the ability to regenerate all neuronal subtypes. This regeneration requires Müller glia (MG) to reprogram and divide asymmetrically to produce a multipotent Müller glia-derived neuronal progenitor cell (MGPC). This raises three key questions. First, does loss of different retinal cell subtypes induce unique MG regeneration responses? Second, do MG reprogram to a developmental retinal progenitor cell state? And finally, to what extent does regeneration recapitulate retinal development? We examined these questions by performing single-nuclear and single-cell RNA-Seq and ATAC-Seq in both developing and regenerating retinas. While MG reprogram to a state similar to late-stage retinal progenitors in developing retinas, there are transcriptional differences between reprogrammed MG/MGPCs and late progenitors, as well as reprogrammed MG in outer and inner retinal damage models. Validation of candidate genes confirmed that loss of different subtypes induces differences in transcription factor gene expression and regeneration outcomes. This work identifies major differences between gene regulatory networks activated following the selective loss of different subtypes of retina neurons, as well as between retinal regeneration and development.

Common and divergent gene regulatory networks control injury-induced and developmental neurogenesis in zebrafish retina.

Following acute retinal damage, zebrafish possess the ability to regenerate all neuronal subtypes. This regeneration requires Müller glia (MG) to reprogram and divide asymmetrically to produce a multipotent Müller glia-derived neuronal progenitor cell (MGPC). This raises three key questions. First, does loss of different retinal cell subtypes induce unique MG regeneration responses? Second, do MG reprogram to a developmental retinal progenitor cell state? And finally, to what extent does regeneration recapitulate retinal development? We examined these questions by performing single-nuclear and single-cell RNA-Seq and ATAC-Seq in both developing and regenerating retinas. While MG reprogram to a state similar to late-stage retinal progenitors in developing retinas, there are transcriptional differences between reprogrammed MG/MGPCs and late progenitors, as well as reprogrammed MG in outer and inner retinal damage models. Validation of candidate genes confirmed that loss of different subtypes induces differences in transcription factor gene expression and regeneration outcomes. This work identifies major differences between gene regulatory networks activated following the selective loss of different subtypes of retina neurons, as well as between retinal regeneration and development.

Photocatalysis of Metallic Nanoparticles: Interband vs Intraband Induced Mechanisms.

Photocatalysis induced by localized surface plasmon resonance of metallic nanoparticles has been studied for more than a decade, but photocatalysis originating from direct interband excitations is still under-explored. The spectral overlap and the coupling of these two optical regimes also complicate the determination of hot carriers' energy states and eventually hinder the accurate assignment of their catalytic role in studied reactions. In this Featured Article, after reviewing previous studies, we suggest classifying the photoexcitation via intra- and interband transitions where the physical states of hot carriers are well-defined. Intraband transitions are featured by creating hot electrons above the Fermi level and suitable for reductive catalytic pathways, whereas interband transitions are featured by generating hot d-band holes below the Fermi level and better for oxidative catalytic pathways. Since the contribution of intra- and interband transitions are different in the spectral regions of localized surface plasmon resonance and direct interband excitations, the wavelength dependence of the photocatalytic activities is very helpful in assigning which transitions and carriers contribute to the observed catalysis.

CellAnn: a comprehensive, super-fast, and user-friendly single-cell annotation web server.

MOTIVATION: Single-cell sequencing technology has become a routine in studying many biological problems. A core step of analyzing single-cell data is the assignment of cell clusters to specific cell types. Reference-based methods are proposed for predicting cell types for single-cell clusters. However, the scalability and lack of preprocessed reference datasets prevent them from being practical and easy to use. RESULTS: Here, we introduce a reference-based cell annotation web server, CellAnn, which is super-fast and easy to use. CellAnn contains a comprehensive reference database with 204 human and 191 mouse single-cell datasets. These reference datasets cover 32 organs. Furthermore, we developed a cluster-to-cluster alignment method to transfer cell labels from the reference to the query datasets, which is superior to the existing methods with higher accuracy and higher scalability. Finally, CellAnn is an online tool that integrates all the procedures in cell annotation, including reference searching, transferring cell labels, visualizing results, and harmonizing cell annotation labels. Through the user-friendly interface, users can identify the best annotation by cross-validating with multiple reference datasets. We believe that CellAnn can greatly facilitate single-cell sequencing data analysis. AVAILABILITY AND IMPLEMENTATION: The web server is available at www.cellann.io, and the source code is available at https://github.com/Pinlyu3/CellAnn_shinyapp.

Gate-Tunable Renormalization of Spin-Correlated Flat-Band States and Bandgap in a 2D Magnetic Insulator.

Emergent quantum phenomena in two-dimensional van der Waal (vdW) magnets are largely governed by the interplay between exchange and Coulomb interactions. The ability to precisely tune the Coulomb interaction enables the control of spin-correlated flat-band states, band gap, and unconventional magnetism in such strongly correlated materials. Here, we demonstrate a gate-tunable renormalization of spin-correlated flat-band states and bandgap in magnetic chromium tribromide (CrBr3) monolayers grown on graphene. Our gate-dependent scanning tunneling spectroscopy (STS) studies reveal that the interflat-band spacing and bandgap of CrBr3 can be continuously tuned by 120 and 240 meV, respectively, via electrostatic injection of carriers into the hybrid CrBr3/graphene system. This can be attributed to the self-screening of CrBr3 arising from the gate-induced carriers injected into CrBr3, which dominates over the weakened remote screening of the graphene substrate due to the decreased carrier density in graphene. Precise tuning of the spin-correlated flat-band states and bandgap in 2D magnets via electrostatic modulation of Coulomb interactions not only provides effective strategies for optimizing the spin transport channels but also may exert a crucial influence on the exchange energy and spin-wave gap, which could raise the critical temperature for magnetic order.

Plasmonic Cavity-Catalysis by Standing Hot Carrier Waves.

Manipulating active sites of catalysts is crucial but challenging in catalysis science and engineering. Beyond the design of the composition and structure of catalysts, the confined electromagnetic field in optical cavities has recently become a promising method for catalyzing chemical reactions via strong light-matter interactions. Another form of confined electromagnetic field, the charge density wave in plasmonic cavities, however, still needs to be explored for catalysis. Here, we present an unprecedented catalytic mode based on plasmonic cavities, called plasmonic cavity-catalysis. We achieve direct control of catalytic sites in plasmonic cavities through standing hot carrier waves. Periodic catalytic hotspots are formed because of localized energy and carrier distribution and can be well tuned by cavity geometry, charge density, and excitation angle. We also found that the catalytic activity of the cavity mode increases several orders of magnitude compared with conventional plasmonic catalysis. We ultimately demonstrate that the locally concentrated long-lived hot carriers in the standing wave mode underlie the formation of the catalytic hotspots. Plasmonic cavity-catalysis provides a new approach to manipulate the catalytic sites and rates and may expand the frontier of heterogeneous catalysis.

Atomically precise vacancy-assembled quantum antidots.

Patterning antidots, which are regions of potential hills that repel electrons, into well-defined antidot lattices creates fascinating artificial periodic structures, leading to anomalous transport properties and exotic quantum phenomena in two-dimensional systems. Although nanolithography has brought conventional antidots from the semiclassical regime to the quantum regime, achieving precise control over the size of each antidot and its spatial period at the atomic scale has remained challenging. However, attaining such control opens the door to a new paradigm, enabling the creation of quantum antidots with discrete quantum hole states, which, in turn, offer a fertile platform to explore novel quantum phenomena and hot electron dynamics in previously inaccessible regimes. Here we report an atomically precise bottom-up fabrication of a series of atomic-scale quantum antidots through a thermal-induced assembly of a chalcogenide single vacancy in PtTe2. Such quantum antidots consist of highly ordered single-vacancy lattices, spaced by a single Te atom, reaching the ultimate downscaling limit of antidot lattices. Increasing the number of single vacancies in quantum antidots strengthens the cumulative repulsive potential and consequently enhances the collective interference of multiple-pocket scattered quasiparticles inside quantum antidots, creating multilevel quantum hole states with a tunable gap from the telecom to far-infrared regime. Moreover, precisely engineered quantum hole states of quantum antidots are geometry protected and thus survive on oxygen substitutional doping. Therefore, single-vacancy-assembled quantum antidots exhibit unprecedented robustness and property tunability, positioning them as highly promising candidates for advancing quantum information and photocatalysis technologies.

Targeting lncRNA NEAT1 Hampers Alzheimer's Disease Progression.

Long noncoding RNA nuclear enriched abundant transcript 1 (lnc-NEAT1) is closely implicated in neurological diseases, while its implication in Alzheimer's disease (AD) is rarely reported. This study aimed to investigate the effect of lnc-NEAT1 knockdown on neuron injury, inflammation, and oxidative stress in AD, as well as its interaction with downstream targets and pathways. APPswe/PS1dE9 transgenic mice were injected with negative control or lnc-NEAT1 interference lentivirus. Besides, AD cellular model was constructed by amyloid ß treatment in mice primary neuron cells; then, knockdown of lnc-NEAT1 and microRNA-193a was performed alone or in combination. In vivo experiments revealed that Lnc-NEAT1 knockdown improved cognition in AD mice reflected by Morrison water maze and Y-maze assays. Besides, lnc-NEAT1 knockdown reduced injury and apoptosis, decreased inflammatory cytokine levels, repressed oxidative stress level, and activated adenosine cyclophosphate response element-binding protein (CREB)/brain-derived neurotrophic factor (BDNF) and nuclear factor erythroid 2-related factor 2 (NRF2)/nicotinamide adenine dinucleotide phosphate dehydrogenase 1 (NQO1) pathways in hippocampi of AD mice. Notably, lnc-NEAT1 down-regulated microRNA-193a both in vitro and in vivo and acted as a decoy of microRNA-193a. In vitro experiments showed that lnc-NEAT1 knockdown decreased apoptosis and oxidative stress, improved cell viability, also activated CREB/BDNF and NRF2/NQO1 pathways in AD cellular model. Meanwhile, microRNA-193a knockdown showed the opposite effects, which also attenuated lnc-NEAT1 knockdown-mediated reduction in injury, oxidative stress, and CREB/BDNF and NRF2/NQO1 pathways of AD cellular model. In conclusion, lnc-NEAT1 knockdown reduces neuron injury, inflammation, and oxidative stress through activating microRNA-193a mediated CREB/BDNF and NRF2/NQO1 pathways in AD.

Ferromagnetic single-atom spin catalyst for boosting water splitting.

Heterogeneous single-atom spin catalysts combined with magnetic fields provide a powerful means for accelerating chemical reactions with enhanced metal utilization and reaction efficiency. However, designing these catalysts remains challenging due to the need for a high density of atomically dispersed active sites with a short-range quantum spin exchange interaction and long-range ferromagnetic ordering. Here, we devised a scalable hydrothermal approach involving an operando acidic environment for synthesizing various single-atom spin catalysts with widely tunable substitutional magnetic atoms (M1) in a MoS2 host. Among all the M1/MoS2 species, Ni1/MoS2 adopts a distorted tetragonal structure that prompts both ferromagnetic coupling to nearby S atoms as well as adjacent Ni1 sites, resulting in global room-temperature ferromagnetism. Such coupling benefits spin-selective charge transfer in oxygen evolution reactions to produce triplet O2. Furthermore, a mild magnetic field of ~0.5 T enhances the oxygen evolution reaction magnetocurrent by ~2,880% over Ni1/MoS2, leading to excellent activity and stability in both seawater and pure water splitting cells. As supported by operando characterizations and theoretical calculations, a great magnetic-field-enhanced oxygen evolution reaction performance over Ni1/MoS2 is attributed to a field-induced spin alignment and spin density optimization over S active sites arising from field-regulated S(p)-Ni(d) hybridization, which in turn optimizes the adsorption energies for radical intermediates to reduce overall reaction barriers.

Catalytically active atomically thin cuprate with periodic Cu single sites.

Rational design and synthesis of catalytically active two-dimensional (2D) materials with an abundance of atomically precise active sites in their basal planes remains a great challenge. Here, we report a ligand exchange strategy to exfoliate bulk [Cu4(OH)6][O3S(CH2)4SO3] cuprate crystals into atomically thin 2D cuprate layers ([Cu2(OH)3]+). The basal plane of 2D cuprate layers contains periodic arrays of accessible unsaturated Cu(II) single sites (2D-CuSSs), which are found to promote efficient oxidative Chan-Lam coupling. Our mechanistic studies reveal that the reactions proceed via coordinatively unsaturated CuO4(II) single sites with the formation of Cu(I) species in the rate-limiting step, as corroborated by both operando experimental and theoretical studies. The robust stability of 2D-CuSSs in both batch and continuous flow reactions, coupled with their recyclability and good performance in complex molecule derivatization, render 2D-CuSSs attractive catalyst candidates for broad utility in fine chemical synthesis.

[Effect of continuous renal replacement therapy on plasma concentration, clinical efficacy and safety of colistin sulfate].

OBJECTIVE: To investigate the effects of continuous renal replacement therapy (CRRT) on plasma concentration, clinical efficacy and safety of colistin sulfate. METHODS: Clinical data of patients received with colistin sulfate were retrospectively analyzed from our group's previous clinical registration study, which was a prospective, multicenter observation study on the efficacy and pharmacokinetic characteristics of colistin sulfate in patients with severe infection in intensive care unit (ICU). According to whether patients received blood purification treatment, they were divided into CRRT group and non-CRRT group. Baseline data (gender, age, whether complicated with diabetes, chronic nervous system disease, etc), general data (infection of pathogens and sites, steady-state trough concentration, steady-state peak concentration, clinical efficacy, 28-day all-cause mortality, etc) and adverse event (renal injury, nervous system, skin pigmentation, etc) were collected from the two groups. RESULTS: A total of 90 patients were enrolled, including 22 patients in the CRRT group and 68 patients in the non-CRRT group. (1) There was no significant difference in gender, age, basic diseases, liver function, infection of pathogens and sites, colistin sulfate dose between the two groups. Compared with the non-CRRT group, the acute physiology and chronic health evaluation II (APACHE II) and sequential organ failure assessment (SOFA) were higher in the CRRT group [APACHE II: 21.77±8.26 vs. 18.01±6.34, P < 0.05; SOFA: 8.5 (7.8, 11.0) vs. 6.0 (4.0, 9.0), P < 0.01], serum creatinine level was higher [µmol/L: 162.0 (119.5, 210.5) vs. 72.0 (52.0, 117.0), P < 0.01]. (2) Plasma concentration: there was no significant difference in steady-state trough concentration between CRRT group and non-CRRT group (mg/L: 0.58±0.30 vs. 0.64±0.25, P = 0.328), nor was there significant difference in steady-state peak concentration (mg/L: 1.02±0.37 vs. 1.18±0.45, P = 0.133). (3) Clinical efficacy: there was no significant difference in clinical response rate between CRRT group and non-CRRT group [68.2% (15/22) vs. 80.9% (55/68), P = 0.213]. (4) Safety: acute kidney injury occurred in 2 patients (2.9%) in the non-CRRT group. No obvious neurological symptoms and skin pigmentation were found in the two groups. CONCLUSIONS: CRRT had little effect on the elimination of colistin sulfate. Routine blood concentration monitoring (TDM) is warranted in patients received with CRRT.

Periodic Distributions and Ultrafast Dynamics of Hot Electrons in Plasmonic Resonators.

Understanding and managing hot electrons in metals are of fundamental and practical interest in plasmonic studies and applications. A major challenge for the development of hot electron devices requires the efficient and controllable generation of long-lived hot electrons so that they can be harnessed effectively before relaxation. Here, we report the ultrafast spatiotemporal evolution of hot electrons in plasmonic resonators. Using femtosecond-resolution interferometric imaging, we show the unique periodic distributions of hot electrons due to standing plasmonic waves. In particular, this distribution can be flexibly tuned by the size, shape, and dimension of the resonator. We also demonstrate that the hot electron lifetimes are substantially prolonged at hot spots. This appealing effect is interpreted as a result of the locally concentrated energy density at the antinodes in standing hot electron waves. These results could be useful to control the distributions and lifetimes of hot electrons in plasmonic devices for targeted optoelectronic applications.

Experimental and Thermodynamic Viewpoints on Claims of a Spontaneous H2O2 Formation at the Air-Water Interface.

Recent claims of the spontaneous H2O2 formation at the air-water interface of water microdroplets have sparked debates on its feasibility. New results from different research groups have provided more insight into these claims, but conclusive proofs are still far from realized. In this Perspective, thermodynamic viewpoints, potential experiments, and theoretical approaches are presented as references for future studies. We suggest that future work should seek for H2 byproduct as indirect evidence to confirm the feasibility of this phenomenon. Examining potential energy surfaces for H2O2 formation reaction when moving from the bulk to the interface under the influence of the local electric fields is also critical to establish this phenomenon.

URINARY METABOLOMICS OF CENTRAL SEROUS CHORIORETINOPATHY.

PURPOSE: To analyze the urinary metabolomic profile of central serous chorioretinopathy cases. METHODS: In a cross-sectional study, 80 participants with central serous chorioretinopathy were compared with 80 age-matched and sex-matched controls. Urinary metabolites were measured using Metabolon's Discovery HD4 platform. RESULTS: Of 1,031 metabolites total that were measured in urine samples, 53 were upregulated and 27 downregulated in central serous chorioretinopathy participants compared with controls. After exclusion of potentially confounding xenobiotics and bile compounds that could represent digestive processes, 14 metabolites were significantly higher and 12 metabolites were significantly lower in cases compared with controls. One upregulated metabolite (tetrahydrocortisol sulfate) is involved in the corticosteroid subpathway. The downregulated metabolites are unrelated to the identified corticosteroid subpathway. CONCLUSION: The upregulation of urinary tetrahydrocortisol sulfate in central serous chorioretinopathy cases provides a precise molecular basis to further study the role of corticosteroids in producing choroidal venous congestion.

Improving performance of deep learning models using 3.5D U-Net via majority voting for tooth segmentation on cone beam computed tomography.

Deep learning allows automatic segmentation of teeth on cone beam computed tomography (CBCT). However, the segmentation performance of deep learning varies among different training strategies. Our aim was to propose a 3.5D U-Net to improve the performance of the U-Net in segmenting teeth on CBCT. This study retrospectively enrolled 24 patients who received CBCT. Five U-Nets, including 2Da U-Net, 2Dc U-Net, 2Ds U-Net, 2.5Da U-Net, 3D U-Net, were trained to segment the teeth. Four additional U-Nets, including 2.5Dv U-Net, 3.5Dv5 U-Net, 3.5Dv4 U-Net, and 3.5Dv3 U-Net, were obtained using majority voting. Mathematical morphology operations including erosion and dilation (E&D) were applied to remove diminutive noise speckles. Segmentation performance was evaluated by fourfold cross validation using Dice similarity coefficient (DSC), accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV). Kruskal-Wallis test with post hoc analysis using Bonferroni correction was used for group comparison. P < 0.05 was considered statistically significant. Performance of U-Nets significantly varies among different training strategies for teeth segmentation on CBCT (P < 0.05). The 3.5Dv5 U-Net and 2.5Dv U-Net showed DSC and PPV significantly higher than any of five originally trained U-Nets (all P < 0.05). E&D significantly improved the DSC, accuracy, specificity, and PPV (all P < 0.005). The 3.5Dv5 U-Net achieved highest DSC and accuracy among all U-Nets. The segmentation performance of the U-Net can be improved by majority voting and E&D. Overall speaking, the 3.5Dv5 U-Net achieved the best segmentation performance among all U-Nets.