HyGAnno: hybrid graph neural network-based cell type annotation for single-cell ATAC sequencing data.

Reliable cell type annotations are crucial for investigating cellular heterogeneity in single-cell omics data. Although various computational approaches have been proposed for single-cell RNA sequencing (scRNA-seq) annotation, high-quality cell labels are still lacking in single-cell sequencing assay for transposase-accessible chromatin (scATAC-seq) data, because of extreme sparsity and inconsistent chromatin accessibility between datasets. Here, we present a novel automated cell annotation method that transfers cell type information from a well-labeled scRNA-seq reference to an unlabeled scATAC-seq target, via a parallel graph neural network, in a semi-supervised manner. Unlike existing methods that utilize only gene expression or gene activity features, HyGAnno leverages genome-wide accessibility peak features to facilitate the training process. In addition, HyGAnno reconstructs a reference-target cell graph to detect cells with low prediction reliability, according to their specific graph connectivity patterns. HyGAnno was assessed across various datasets, showcasing its strengths in precise cell annotation, generating interpretable cell embeddings, robustness to noisy reference data and adaptability to tumor tissues.

iSMOD: an integrative browser for image-based single-cell multi-omics data.

Genomic and transcriptomic image data, represented by DNA and RNA fluorescence in situ hybridization (FISH), respectively, together with proteomic data, particularly that related to nuclear proteins, can help elucidate gene regulation in relation to the spatial positions of chromatins, messenger RNAs, and key proteins. However, methods for image-based multi-omics data collection and analysis are lacking. To this end, we aimed to develop the first integrative browser called iSMOD (image-based Single-cell Multi-omics Database) to collect and browse comprehensive FISH and nucleus proteomics data based on the title, abstract, and related experimental figures, which integrates multi-omics studies focusing on the key players in the cell nucleus from 20 000+ (still growing) published papers. We have also provided several exemplar demonstrations to show iSMOD's wide applications-profiling multi-omics research to reveal the molecular target for diseases; exploring the working mechanism behind biological phenomena using multi-omics interactions, and integrating the 3D multi-omics data in a virtual cell nucleus. iSMOD is a cornerstone for delineating a global view of relevant research to enable the integration of scattered data and thus provides new insights regarding the missing components of molecular pathway mechanisms and facilitates improved and efficient scientific research.

Telecom-wavelength conversion in a high optical depth cold atomic system.

We experimentally investigate the frequency down-conversion through the four-wave mixing (FWM) process in a cold 85Rb atomic ensemble, with a diamond-level configuration. An atomic cloud with a high optical depth (OD) of 190 is prepared to achieve a high efficiency frequency conversion. Here, we convert a signal pulse field (795 nm) attenuated to a single-photon level, into a telecom light at 1529.3 nm within near C-band range and the frequency-conversion efficiency can reach up to 32%. We find that the OD is an essential factor affecting conversion efficiency and the efficiency may exceed 32% with an improvement in the OD. Moreover, we note the signal-to-noise ratio of the detected telecom field is higher than 10 while the mean signal count is larger than 0.2. Our work may be combined with quantum memories based on cold 85Rb ensemble at 795 nm and serve for long-distance quantum networks.

Optical memory for arbitrary perfect Poincaré states in an atomic ensemble.

Inherent spin angular momentum (SAM) and orbital angular momentum (OAM), which manifest as polarization and spatial degrees of freedom (DOFs) of photons, hold a promise of large capability for applications in classical and quantum information processing. To enable these photonic spin and orbital dynamic properties strongly coupled with each other, Poincaré states have been proposed and offer advantages in data multiplexing, information encryption, precision metrology, and quantum memory. However, since the transverse size of Laguerre-Gaussian beams strongly depends on their topological charge numbers | l |, it is difficult to store asymmetric Poincaré states due to the significantly different light-matter interaction for distinct spatial modes. Here, we experimentally realize the storage of perfect Poincaré states with arbitrary OAM quanta using the perfect optical vortex, in which 121 arbitrarily selected perfect Poincaré states have been stored with high fidelity. The reported work has great prospects in optical communication and quantum networks for dramatically increased encoding flexibility of information.

Highly Efficient Storage of 25-Dimensional Photonic Qudit in a Cold-Atom-Based Quantum Memory.

Building an efficient quantum memory in high-dimensional Hilbert spaces is one of the fundamental requirements for establishing high-dimensional quantum repeaters, where it offers many advantages over two-dimensional quantum systems, such as a larger information capacity and enhanced noise resilience. To date, it remains a challenge to develop an efficient high-dimensional quantum memory. Here, we experimentally realize a quantum memory that is operational in Hilbert spaces of up to 25 dimensions with a storage efficiency of close to 60% and a fidelity of 84.2±0.6%. The proposed approach exploits the spatial-mode-independent interaction between atoms and photons which are encoded in transverse-size-invariant vortex modes. In particular, our memory features uniform storage efficiency and low crosstalk disturbance for 25 individual spatial modes of photons, thus allowing the storing of qudit states programmed from 25 eigenstates within the high-dimensional Hilbert spaces. These results have great prospects for the implementation of long-distance high-dimensional quantum networks and quantum information processing.

Liquid nitric oxide donor for adjuvant therapy of acute ischemic stroke via nasal administration.

OBJECTIVES: The aim of this study was to synthesize and characterize a novel NO donor, PEI-PO-NONOate, using propylene oxide and to investigate its biosafety and therapeutic efficacy via nasal administration in vitro and vivo. EXPERIMENTAL PROCEDURES: The PEI-PO-NONOate was synthesized based on polyethylenimine (PEI) with different molecular weights and characterized using Fourier transform infrared (FTIR), nuclear magnetic resonance (NMR), and ultraviolet (UV) spectroscopy. Cytotoxicity assays were performed on mouse fibroblast cells L929 and human nasal mucosa epithelial cells (HNEpC), and a rat middle cerebral artery occlusion (MCAO) model was established to evaluate the therapeutic efficacy of PEI-PO-NONOate via nasal administration. RESULTS: The PEI-PO-NONOate was found to be stable under dark, dry, and airproof conditions, and its release was accelerated in an aqueous phase or acidic environment, while it was slowed down in a polyethylene glycol (PEG) mixture system. The NO donor released approximately 0.4, 0.5, and 0.6 µmol of gaseous NO from 1.0 mg of the polymer based on PEI600, PEI1800, and PEI10K, respectively. Cytotoxicity assays showed that the PEI-PO-NONOates had a cryoprotective effect as compared with PEI and PEI-PO. Furthermore, nasal administration of PEI-PO-NONOates resulted in a significant reduction in overall necrotic ratio as compared with the control group (16.4% versus 24.6%, p < 0.05). CONCLUSION: The findings of this study suggest that PEI-PO-NONOates may have potential as an adjuvant therapy for acute ischemic stroke when administered via the nasal route.

Accurate identification and measurement of the precipitate area by two-stage deep neural networks in novel chromium-based alloys.

The performance of advanced materials for extreme environments is underpinned by their microstructure, such as the size and distribution of nano- to micro-sized reinforcing phase(s). Chromium-based superalloys are a recently proposed alternative to conventional face-centred-cubic superalloys for high-temperature applications, e.g., Concentrated Solar Power. Their development requires the determination of precipitate volume fraction and size distribution using Electron Microscopy (EM), as these properties are crucial for the thermal stability and mechanical properties of chromium superalloys. Traditional approaches to EM image processing utilise filtering with a fixed contrast threshold, leads to weak robustness to background noise and poor generalisability to different materials. It also requires an enormous amount of time for manual object measurements on large datasets. Efficient and accurate object detection and segmentation are therefore highly desired to accelerate the development of novel materials like chromium-based superalloys. To address these bottlenecks, based on YOLOv5 and SegFormer structures, this study proposes an end-to-end, two-stage deep learning scheme, DT-SegNet, to perform object detection and segmentation for EM images. The proposed approach can thus benefit from the training efficiency of CNNs at the detection stage (i.e., a small number of training images required) and the accuracy of the ViT at the segmentation stage. Extensive numerical experiments demonstrate that the proposed DT-SegNet significantly outperforms the state-of-the-art segmentation tools offered by Weka and ilastik regarding a large number of metrics, including accuracy, precision, recall and F1-score. This model forms a useful tool to aid alloy development microstructure examinations, and offers significant advantages to address the large datasets associated with high-throughput alloy development approaches.

Effects of "fixation-fusion" sequence of lumbar surgery on surgical outcomes for patients with lumbar spinal stenosis: study protocol for a multicenter randomized controlled trial.

BACKGROUND: New-onset neurological symptoms such as numbness and pain in lower extremities might appear immediately after conventional lumbar interbody fusion (LIF) surgery performed in patients with lumbar spinal stenosis. METHODS AND ANALYSIS: This is a multicenter, randomized, open-label, parallel-group, active-controlled trial investigating the clinical outcomes of modified LIF sequence versus conventional LIF sequence in treating patients with lumbar spinal stenosis. A total of 254 eligible patients will be enrolled and randomized in a 1:1 ratio to either modified LIF sequence or conventional LIF sequence group. The primary outcome measure is the perioperative incidence of new-onset lower extremity neurological symptoms, including new adverse events of pain, numbness, and foot drop of any severity. Important secondary endpoints include visual analogue scale (VAS) pain score and lumbar Japanese Orthopaedic Association (JOA) recovery rate. Other safety endpoints will also be evaluated. The safety set used for safety data analysis by the actual surgical treatment received and the full analysis set for baseline and efficacy data analyses according to the intent-to-treat principle will be established as the two analysis populations in the study. CONCLUSION: This study is designed to investigate the clinical outcomes of modified LIF sequences in patients with lumbar spinal stenosis. It aims to provide clinical evidence that the modified "fixation-fusion" sequence of LIF surgery is effective in treating lumbar spinal stenosis. TRIAL REGISTRATION: http://www.chictr.org.cn/index.aspx ID: ChiCTR2100048507.

Long-Lived Memory for Orbital Angular Momentum Quantum States.

Quantum memories that are capable of storing multiple spatial modes offer advantages in speed and robustness when incorporated into quantum networks. When it comes to spatial degrees of freedom, orbital angular momentum (OAM) modes have received widespread attention since they enable encoding with inherent infinite number of dimensions. Although the faithful storage of OAM qubits or qutrits has been realized in previous works, the achieved lifetimes are still on the order of a few microseconds as limited by the spatially dependent decoherence. We here demonstrate a long-lived quantum memory for OAM qutrits by suppressing the decoherence in the transverse and longitude direction simultaneously; the achieved fidelity beats the quantum-classical criteria after a storage time of 400 µs, which is 2 orders of magnitude longer than earlier works. The present work is promising for establishing high-dimensional quantum networks.

High-axial-resolution single-molecule localization under dense excitation with a multi-channel deep U-Net.

Single-molecule localization microscopy (SMLM) can bypass the diffraction limit of optical microscopes and greatly improve the resolution in fluorescence microscopy. By introducing the point spread function (PSF) engineering technique, we can customize depth varying PSF to achieve higher axial resolution. However, most existing 3D single-molecule localization algorithms require excited fluorescent molecules to be sparse and captured at high signal-to-noise ratios, which results in a long acquisition time and precludes SMLM's further applications in many potential fields. To address this problem, we propose a novel 3D single-molecular localization method based on a multi-channel neural network based on U-Net. By leveraging the deep network's great advantages in feature extraction, the proposed network can reliably discriminate dense fluorescent molecules with overlapped PSFs and corrupted by sensor noise. Both simulated and real experiments demonstrate its superior performance in PSF engineered microscopes with short exposure and dense excitations, which holds great potential in fast 3D super-resolution microscopy.

HOPE: High-Order Polynomial Expansion of Black-Box Neural Networks.

Despite their remarkable performance, deep neural networks remain mostly "black boxes", suggesting inexplicability and hindering their wide applications in fields requiring making rational decisions. Here we introduce HOPE  (High-order Polynomial Expansion), a method for expanding a network into a high-order Taylor polynomial on a reference input. Specifically, we derive the high-order derivative rule for composite functions and extend the rule to neural networks to obtain their high-order derivatives quickly and accurately. From these derivatives, we can then derive the Taylor polynomial of the neural network, which provides an explicit expression of the network's local interpretations. We combine the Taylor polynomials obtained under different reference inputs to obtain the global interpretation of the neural network. Numerical analysis confirms the high accuracy, low computational complexity, and good convergence of the proposed method. Moreover, we demonstrate HOPE's wide applications built on deep learning, including function discovery, fast inference, and feature selection. We compared HOPE  with other XAI methods and demonstrated our advantages. The code is available at https://github.com/HarryPotterXTX/HOPE.git.

Encoder-Decoder Contrast for Unsupervised Anomaly Detection in Medical Images.

Unsupervised anomaly detection (UAD) aims to recognize anomalous images based on the training set that contains only normal images. In medical image analysis, UAD benefits from leveraging the easily obtained normal (healthy) images, avoiding the costly collecting and labeling of anomalous (unhealthy) images. Most advanced UAD methods rely on frozen encoder networks pre-trained using ImageNet for extracting feature representations. However, the features extracted from the frozen encoders that are borrowed from natural image domains coincide little with the features required in the target medical image domain. Moreover, optimizing encoders usually causes pattern collapse in UAD. In this paper, we propose a novel UAD method, namely Encoder-Decoder Contrast (EDC), which optimizes the entire network to reduce biases towards pre-trained image domain and orient the network in the target medical domain. We start from feature reconstruction approach that detects anomalies from reconstruction errors. Essentially, a contrastive learning paradigm is introduced to tackle the problem of pattern collapsing while optimizing the encoder and the reconstruction decoder simultaneously. In addition, to prevent instability and further improve performances, we propose to bring globality into the contrastive objective function. Extensive experiments are conducted across four medical image modalities including optical coherence tomography, color fundus image, brain MRI, and skin lesion image, where our method outperforms all current state-of-the-art UAD methods. Code is available at: https://github.com/guojiajeremy/EDC.

Anomaly Detection for Medical Images Using Heterogeneous Auto-Encoder.

Anomaly detection is an important task for medical image analysis, which can alleviate the reliance of supervised methods on large labelled datasets. Most existing methods use a pixel-wise self-reconstruction framework for anomaly detection. However, there are two challenges of these studies: 1) they tend to overfit learning an identity mapping between the input and output, which leads to failure in detecting abnormal samples; 2) the reconstruction considers the pixel-wise differences which may lead to an undesirable result. To mitigate the above problems, we propose a novel heterogeneous Auto-Encoder (Hetero-AE) for medical anomaly detection. Our model utilizes a convolutional neural network (CNN) as the encoder and a hybrid CNN-Transformer network as the decoder. The heterogeneous structure enables the model to learn the intrinsic information of normal data and enlarge the difference on abnormal samples. To fully exploit the effectiveness of Transformer in the hybrid network, a multi-scale sparse Transformer block is proposed to trade off modelling long-range feature dependencies and high computational costs. Moreover, the multi-stage feature comparison is introduced to reduce the noise of pixel-wise comparison. Extensive experiments on four public datasets (i.e., retinal OCT, chest X-ray, brain MRI, and COVID-19) verify the effectiveness of our method on different imaging modalities for anomaly detection. Additionally, our method can accurately detect tumors in brain MRI and lesions in retinal OCT with interpretable heatmaps to locate lesion areas, assisting clinicians in diagnosing abnormalities efficiently.

PatchCL-AE: Anomaly detection for medical images using patch-wise contrastive learning-based auto-encoder.

Anomaly detection is an important yet challenging task in medical image analysis. Most anomaly detection methods are based on reconstruction, but the performance of reconstruction-based methods is limited due to over-reliance on pixel-level losses. To address the limitation, we propose a patch-wise contrastive learning-based auto-encoder for medical anomaly detection. The key contribution is the patch-wise contrastive learning loss that provides supervision on local semantics to enforce semantic consistency between corresponding input-output patches. Contrastive learning pulls corresponding patch pairs closer while pushing non-corresponding ones apart between input and output, enabling the model to learn local normal features better and improve discriminability on anomalous regions. Additionally, we design an anomaly score based on local semantic discrepancies to pinpoint abnormalities by comparing feature difference rather than pixel variations. Extensive experiments on three public datasets (i.e., brain MRI, retinal OCT, and chest X-ray) achieve state-of-the-art performance, with our method achieving over 99% AUC on retinal and brain images. Both the contrastive patch-wise supervision and patch-discrepancy score provide targeted advancements to overcome the weaknesses in existing approaches.

Multi-omics clustering for cancer subtyping based on latent subspace learning.

The increased availability of high-throughput technologies has enabled biomedical researchers to learn about disease etiology across multiple omics layers, which shows promise for improving cancer subtype identification. Many computational methods have been developed to perform clustering on multi-omics data, however, only a few of them are applicable for partial multi-omics in which some samples lack data in some types of omics. In this study, we propose a novel multi-omics clustering method based on latent sub-space learning (MCLS), which can deal with the missing multi-omics for clustering. We utilize the data with complete omics to construct a latent subspace using PCA-based feature extraction and singular value decomposition (SVD). The data with incomplete multi-omics are then projected to the latent subspace, and spectral clustering is performed to find the clusters. The proposed MCLS method is evaluated on seven different cancer datasets on three levels of omics in both full and partial cases compared to several state-of-the-art methods. The experimental results show that the proposed MCLS method is more efficient and effective than the compared methods for cancer subtype identification in multi-omics data analysis, which provides important references to a comprehensive understanding of cancer and biological mechanisms. AVAILABILITY: The proposed method can be freely accessible at https://github.com/ShangCS/MCLS.

Microbe-mediated simultaneous nitrogen reduction and sulfamethoxazole/N-acetylsulfamethoxazole removal in lab-scale constructed wetlands.

Constructed wetlands (CWs) are increasingly used to treat complex pollution such as nitrogen and emerging organic micropollutants from anthropogenic sources. In this study, the denitrification, anaerobic ammonium oxidation, dissimilatory nitrate reduction to ammonium, and nitrous oxide release rates following exposure to the frequently detected sulfonamides sulfamethoxazole (SMX) and its human metabolite, N-acetylsulfamethoxazole (N-SMX), were investigated in lab-scale CWs. Over a period of 190 d, the denitrification rates were noticeably inhibited in the SMX and N-SMX groups at week 5. Subsequently, the denitrification rates recovered, accompanied by an increase in the relevant nitrogen reduction and antibiotic resistance genes (ARGs). The composition of the microbial community also changed during this process. After the denitrification rates recovered, Burkholderia_Paraburkholderia and Gordonia exhibited a significant positive correlation with SMX exposure, which simultaneously reduced nitrate concentrations and degraded antibiotics. Burkholderia_Paraburkholderia is a key carrier of ARGs. Finally, nitrogen reduction (> 90%) and antibiotic removal (> 80%) also recovered in both SMX- and N-SMX-exposed lab-scale CWs during the operation, which revealed the interaction of SMX or N-SMX removal and nitrogen reduction.

Semi-supervised peripapillary atrophy segmentation with shape constraint.

Peripapillary atrophy (PPA) is a clinical abnormality related to many eye diseases, such as myopia and glaucoma. The shape and area of PPA are essential indicators of disease progression. PPA segmentation is a challenging task due to blurry edge and limited labeled data. In this paper, we propose a novel semi-supervised PPA segmentation method enhanced by prior knowledge. In order to learn shape information in the network, a novel shape constraint module is proposed to restrict the PPA appearance based on active shape model. To further leverage large amount of unlabeled data, a Siamese-like model updated by exponential moving average is introduced to provide pseudo labels. The pseudo labels are further refined by region connectivity correction. Extensive experiments on a clinical dataset demonstrate that our proposed PPA segmentation method provides good qualitative and quantitative performance.

Development and validation of nomogram for the prediction of preterm delivery based on patient characteristics and circulating inflammatory cells in patients with gestational diabetes mellitus.

Background: The incidence of preterm delivery (<37 weeks' gestation) is increased due to gestational diabetes mellitus (GDM). The preterm delivery is the leading cause of death in children. If potential preterm delivery can be diagnosed early and then prevented, adverse pregnancy outcomes can be improved. Therefore, effective methods are needed for early prediction of preterm delivery in women with GDM. Methods: Patients with GDM defined as the presence of at least 1 plasma glucose abnormality at 24-28 weeks of pregnancy [fasting plasma glucose ≥5.1 mmol/L, 60-min ≥10.0 mmol/L, 120-min ≥8.5 mmol/L by 75 g oral glucose tolerance test (OGTT)] from the First Affiliated Hospital of Wenzhou Medical University were enrolled. The data (564 patients) recorded from January 2017 to June 2020 were named the training cohort, and the data (242 patients) obtained from patients with GDM, from July 2020 to January 2022, were named the validation cohort. Mann-Whitney U test and chi-square test were used to compare the skewed distributed and categorical data, respectively. According to the results of univariate logistic regression analysis, the multivariate logistic regression model was developed in the training cohort. Then, the nomogram was established. The validation of the nomogram was conducted on the training and validation cohort. Results: No significant differences in baseline characteristics were detected between the 2 cohorts (all P>0.05). The multivariate analysis suggested that maternal age, insulin use, NLR, and monocyte count were the independent predictors of preterm delivery. A nomogram for predicting the probability of preterm delivery was developed. The model suggested good discrimination [areas under the curve (AUC) =0.885, 95% confidence interval (95% CI): 0.855-0.910, sensitivity =83.0%, specificity =83.1% in the training cohort; AUC =0.919, 95% CI: 0.858-0.980, sensitivity =90.6%, specificity =84.8% in the validation cohort] and good calibration [Hosmer-Lemeshow (HL) test: χ2=3.618, P=0.306 in the training cohort; χ2=6.012, P=0.111 in the validation cohort]. Conclusions: The visual nomogram model appears to be a reliable approach for the prediction of preterm delivery, allowing clinicians to take timely measures to prevent the occurrence of preterm delivery in women with GDM at the time of GDM diagnosis, and deserves further investigation.

Handheld snapshot multi-spectral camera at tens-of-megapixel resolution.

Multi-spectral imaging is a fundamental tool characterizing the constituent energy of scene radiation. However, current multi-spectral video cameras cannot scale up beyond megapixel resolution due to optical constraints and the complexity of the reconstruction algorithms. To circumvent the above issues, we propose a tens-of-megapixel handheld multi-spectral videography approach (THETA), with a proof-of-concept camera achieving 65-megapixel videography of 12 wavebands within visible light range. The high performance is brought by multiple designs: We propose an imaging scheme to fabricate a thin mask for encoding spatio-spectral data using a conventional film camera. Afterwards, a fiber optic plate is introduced for building a compact prototype supporting pixel-wise encoding with a large space-bandwidth product. Finally, a deep-network-based algorithm is adopted for large-scale multi-spectral data decoding, with the coding pattern specially designed to facilitate efficient coarse-to-fine model training. Experimentally, we demonstrate THETA's advantageous and wide applications in outdoor imaging of large macroscopic scenes.

A fundus image classification framework for learning with noisy labels.

Fundus images are widely used in the screening and diagnosis of eye diseases. Current classification algorithms for computer-aided diagnosis in fundus images rely on large amounts of data with reliable labels. However, the appearance of noisy labels degrades the performance of data-dependent algorithms, such as supervised deep learning. A noisy label learning framework suitable for the multiclass classification of fundus diseases is presented in this paper, which combines data cleansing (DC), adaptive negative learning (ANL), and sharpness-aware minimization (SAM) modules. Firstly, the DC module filters the noisy labels in the training dataset based on the prediction confidence. Then, the ANL module modifies the loss function by choosing complementary labels, which are neither the given labels nor the labels with the highest confidence. Moreover, for better generalization, the SAM module is applied by simultaneously optimizing the loss and its sharpness. Extensive experiments on both private and public datasets show that our method greatly promotes the performance for classification of multiple fundus diseases with noisy labels.