Genome-Wide Detection of Quantitative Trait Loci and Prediction of Candidate Genes for Seed Sugar Composition in Early Mature Soybean.

Seed sugar composition, mainly including fructose, glucose, sucrose, raffinose, and stachyose, is an important indicator of soybean [Glycine max (L.) Merr.] seed quality. However, research on soybean sugar composition is limited. To better understand the genetic architecture underlying the sugar composition in soybean seeds, we conducted a genome-wide association study (GWAS) using a population of 323 soybean germplasm accessions which were grown and evaluated under three different environments. A total of 31,245 single-nucleotide polymorphisms (SNPs) with minor allele frequencies (MAFs) ≥ 5% and missing data ≤ 10% were selected and used in the GWAS. The analysis identified 72 quantitative trait loci (QTLs) associated with individual sugars and 14 with total sugar. Ten candidate genes within the 100 Kb flanking regions of the lead SNPs across six chromosomes were significantly associated with sugar contents. According to GO and KEGG classification, eight genes were involved in the sugar metabolism in soybean and showed similar functions in Arabidopsis. The other two, located in known QTL regions associated with sugar composition, may play a role in sugar metabolism in soybean. This study advances our understanding of the genetic basis of soybean sugar composition and facilitates the identification of genes controlling this trait. The identified candidate genes will help improve seed sugar composition in soybean.

Causes and treatment strategies of unilateral leaflet escape of bileaflet mechanical prosthetic heart valves after surgery: a case series.

BACKGROUND: During the eleven years from 2010 to 2021, preliminary statistics have shown that Fuwai Hospital completed 23,571 mechanical valve replacements for various types of valves, and 1139 mechanical valve replacements were performed in Guangyuan First People's Hospital. Only two patients developed valve leaflet escape, so valve leaflet escape is a rare postoperative complication. CASE PRESENTATION: In 2010 and 2021, two patients were selected after they had unilateral leaflet escape after having mechanical valve replacements in Fuwai Hospital of Chinese Academy of Medical Sciences and Guangyuan First People's Hospital. Both patients underwent reoperations with the classic operation and the new bileaflet mechanical prosthetic heart valve was sutured. The treatment of detached single lobe and distal vessel was comprehensively determined, and the condition was treated according to the patient's symptoms, CT results, ultrasound results and other test results, as well as whether this detached lobe caused any abnormal hemodynamics of the distal vessel. The patient with mechanical aortic valve escape completed the 10-year follow-up, and patient with mechanical mitral valve escape completed the 3-month follow-up. there was no thrombosis or hematoma at the embolic site; the patient had no lower limb symptoms. CONCLUSIONS: The reason for the leaflet escape may be related to the valve design and the leaflet material. If the detached leaflets are damaged and if the distal blood vessels are affected, simultaneous surgical treatment is required. Those patients whose vessels were not damaged by the valve lobe should be carefully monitored.

Uncertainty Estimation With Neural Processes for Meta-Continual Learning.

The ability to evaluate uncertainties in evolving data streams has become equally, if not more, crucial than building a static predictor. For instance, during the pandemic, a model should consider possible uncertainties such as governmental policies, meteorological features, and vaccination schedules. Neural process families (NPFs) have recently shone a light on predicting such uncertainties by bridging Gaussian processes (GPs) and neural networks (NNs). Their abilities to output average predictions and the acceptable variances, i.e., uncertainties, made them suitable for predictions with insufficient data, such as meta-learning or few-shot learning. However, existing models have not addressed continual learning which imposes a stricter constraint on the data access. Regarding this, we introduce a member meta-continual learning with neural process (MCLNP) for uncertainty estimation. We enable two levels of uncertainty estimations: the local uncertainty on certain points and the global uncertainty p(z) that represents the function evolution in dynamic environments. To facilitate continual learning, we hypothesize that the previous knowledge can be applied to the current task, hence adopt a coreset as a memory buffer to alleviate catastrophic forgetting. The relationships between the degree of global uncertainties with the intratask diversity and model complexity are discussed. We have estimated prediction uncertainties with multiple evolving types including abrupt/gradual/recurrent shifts. The applications encompass meta-continual learning in the 1-D, 2-D datasets, and a novel spatial-temporal COVID dataset. The results show that our method outperforms the baselines on the likelihood and can rebound quickly even for heavily evolved data streams.

Temporal contrast improvement through cascaded second-order nonlinear processes in a thin BBO crystal.

We demonstrate temporal contrast improvement through cascaded second-order nonlinear processes in a 340-µm BBO crystal. The process was initiated by second harmonic (SH) generation, followed by difference frequency generation (DFG) between the SH and the short wavelength part of the fundamental wave (FW). The idler of DFG was selected by a spectral filter, and an output pulse energy of 573 µJ was obtained at 1 kHz with excellent spatial profile and a power fluctuation as low as 0.076% (rms) in 14 hours. The temporal contrast was improved by more than 2 orders of magnitude to approximately 1011, which could be further enhanced with different spectral filters. The excellent stability, energy scalability, and contrast enhancement ability make this simple and robust method very suitable to be integrated into the pulse cleaning system in many different ultra-intense laser facilities.

[ZHOU Mei-sheng's academic thought and clinical application of "three-phase moxibustion sensation"].

This paper collects professor ZHOU Mei-sheng's academic thought, "three-phases moxibustion sensation" and expounds its clinical value. Proposed by professor ZHOU, in accordance with the occurrence and development characteristics of the moxibustion propagated sensation, three time phases of moxibustion propagated sensation are divided, i.e. directional conduction phase (the first time phase), effect onset phase (the second time phase), and descending suspension and along-meridian re-transmission phase (the third time phase). In terms of the different characteristics among these three time phases, the clinical therapeutic regimens are designed accordingly. It provides a novel approach to the clinical application of moxibustion.

Transcriptomic Analysis Reveals the Regulatory Networks and Hub Genes Controlling the Unsaturated Fatty Acid Contents of Developing Seed in Soybean.

Soybean [Glycine max (L.) Merr.] is one of the most important crops, which produces about 25% of the world's edible oil. The nutritional value of soybean oil depends mostly on the relative contents of three unsaturated fatty acids (UFAs), i.e., oleic acid, linoleic acid (LA), and linolenic acid. However, the biosynthetic mechanism of UFAs remains largely unknown, and there are few studies on RNA-seq analysis of developing seeds. To identify the candidate genes and related pathways involved in the regulation of UFA contents during seed development in soybean, two soybean lines with different UFA profiles were selected from 314 cultivars and landraces originated from Southern China, and RNA-seq analysis was performed in soybean seeds at three developmental stages. Using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, a series of genes and pathways related to fatty acid metabolism were identified, and 40 days after flowering (DAF) was found to be the crucial period in the formation of UFA profiles. Further, weighted gene co-expression network analysis identified three modules with six genes whose functions were highly associated with the contents of oleic and LA. The detailed functional investigation of the networks and hub genes could further improve the understanding of the underlying molecular mechanism of UFA contents and might provide some ideas for the improvement in fatty acids profiles in soybean.

Adversarial Robustness of Deep Reinforcement Learning Based Dynamic Recommender Systems.

Adversarial attacks, e.g., adversarial perturbations of the input and adversarial samples, pose significant challenges to machine learning and deep learning techniques, including interactive recommendation systems. The latent embedding space of those techniques makes adversarial attacks challenging to detect at an early stage. Recent advance in causality shows that counterfactual can also be considered one of the ways to generate the adversarial samples drawn from different distribution as the training samples. We propose to explore adversarial examples and attack agnostic detection on reinforcement learning (RL)-based interactive recommendation systems. We first craft different types of adversarial examples by adding perturbations to the input and intervening on the casual factors. Then, we augment recommendation systems by detecting potential attacks with a deep learning-based classifier based on the crafted data. Finally, we study the attack strength and frequency of adversarial examples and evaluate our model on standard datasets with multiple crafting methods. Our extensive experiments show that most adversarial attacks are effective, and both attack strength and attack frequency impact the attack performance. The strategically-timed attack achieves comparative attack performance with only 1/3 to 1/2 attack frequency. Besides, our white-box detector trained with one crafting method has the generalization ability over several other crafting methods.

Effect of 3D-printed hearts used in left ventricular outflow tract obstruction: a multicenter study.

OBJECTIVE: The purpose of this research was to explore the application value of a three-dimensional (3D)-printed heart in surgery for left ventricular outflow tract (LVOT) obstruction. METHODS: From August 2019 to October 2021, 46 patients with LVOT obstruction underwent surgical treatment at our institution. According to the treatment method, 22 and 24 patients were allocated to the experimental and control groups, respectively. In the experimental group, each patient's 3D-printed heart model was used for simulated preoperative surgery, and then the Morrow operation was performed. In the control group, only the Morrow operation was performed, without simulated preoperative surgery using a 3D-printed heart model. The intraoperative and postoperative data of patients in the two groups were recorded, and the clinical data of patients were compared between the two groups. RESULTS: The operation time, cardiopulmonary bypass time, intraoperative blood loss, hospitalization time, LVOT pressure difference (LVP), postoperative interventricular septal thickness (IST), aortic regurgitation (AR), systolic anterior motion (SAM), and postoperative left ventricular flow velocity (LVFV) were significantly lower in the experimental group than in the control group (P < 0.05). The inner diameter of the left ventricular outflow tract (IDLV) was larger in the experimental group than in the control group (P < 0.05). There was no significant difference in the postoperative ejection fraction, atrioventricular block rate or complication rate between the two groups (P > 0.05). CONCLUSION: A 3D-printed heart model for simulated surgery in vitro is conducive to formulating a more reasonable surgical plan and reducing the trauma and duration of surgery, thereby promoting the recovery and maintenance of the heart.

Industrial gearbox fault diagnosis based on multi-scale convolutional neural networks and thermal imaging.

Infrared thermal technology plays a vital role in the health condition monitoring of gearbox. In the traditional infrared thermal technology-based methods, Gaussian pyramid is applied as the feature extraction approach, which has disadvantages of noise influence and information missing. Focus on such disadvantages, an improved multi-scale decomposition method combined with convolutional neural network is proposed to extract the fault features of the multi-scale infrared images in this paper. It can enlarge the data length at large scales, and thus reduce the fluctuations of feature values and reserve the fault information. The effectiveness of the proposed method is validated using the experiment infrared data of one industrial gearbox. Results demonstrate that our proposed method has the best performance comparing with five methods.

Explainable depression detection with multi-aspect features using a hybrid deep learning model on social media.

The ability to explain why the model produced results in such a way is an important problem, especially in the medical domain. Model explainability is important for building trust by providing insight into the model prediction. However, most existing machine learning methods provide no explainability, which is worrying. For instance, in the task of automatic depression prediction, most machine learning models lead to predictions that are obscure to humans. In this work, we propose explainable Multi-Aspect Depression Detection with Hierarchical Attention Network MDHAN, for automatic detection of depressed users on social media and explain the model prediction. We have considered user posts augmented with additional features from Twitter. Specifically, we encode user posts using two levels of attention mechanisms applied at the tweet-level and word-level, calculate each tweet and words' importance, and capture semantic sequence features from the user timelines (posts). Our hierarchical attention model is developed in such a way that it can capture patterns that leads to explainable results. Our experiments show that MDHAN outperforms several popular and robust baseline methods, demonstrating the effectiveness of combining deep learning with multi-aspect features. We also show that our model helps improve predictive performance when detecting depression in users who are posting messages publicly on social media. MDHAN achieves excellent performance and ensures adequate evidence to explain the prediction.

Agglomerative Neural Networks for Multiview Clustering.

Conventional multiview clustering methods seek a view consensus through minimizing the pairwise discrepancy between the consensus and subviews. However, pairwise comparison cannot portray the interview relationship precisely if some of the subviews can be further agglomerated. To address the above challenge, we propose the agglomerative analysis to approximate the optimal consensus view, thereby describing the subview relationship within a view structure. We present an agglomerative neural network (ANN) based on constrained Laplacian rank to cluster multiview data directly without a dedicated postprocessing step (e.g., using K -means). We further extend ANN with a learnable data space to handle data of complex scenarios. Our evaluations against several state-of-the-art multiview clustering approaches on four popular data sets show the promising view-consensus analysis ability of ANN. We further demonstrate ANN's capability in analyzing complex view structures, extensibility through our case study and robustness and effectiveness of data-driven modifications.

Nutritional evaluation and transcriptome analyses of short-time germinated seeds in soybean (Glycine max L. Merri.).

Germination is a common practice for nutrition improvement in many crops. In soybean, the nutrient value and genome-wide gene expression pattern of whole seeds germinated for short-time has not been fully investigated. In this study, protein content (PC), water soluble protein content (WSPC), isoflavone compositions were evaluated at 0 and 36 h after germination (HAG), respectively. The results showed that at 36HAG, PC was slightly decreased (P > 0.05) in ZD41, J58 and JHD, WSPC and free isoflavone (aglycones: daidzein, genistein, and glycitein) were significantly increased (P < 0.05), while total isoflavone content was unchanged. Transcriptomic analysis identified 5240, 6840 and 15,766 DEGs in different time point comparisons, respectively. GO and KEGG analysis showed that photosynthesis process was significantly activated from 18HAG, and alternative splicing might play an important role during germination in a complex manner. Response to hydrogen peroxide (H2O2) was found to be down regulated significantly from 18 to 36HAG, suggesting that H2O2 might play an important role in germination. Expression pattern analysis showed the synthesis of storage proteins was slowing down, while the genes coding for protein degradation (peptidase and protease) were up regulated as time went by during germination. For genes involved in isoflavone metabolism pathway, UGT (7-O-glucosyltransferase) coding genes were significantly up regulated (40 up-DEGs vs 27 down-DEGs), while MAT (7-O-glucoside-6''-O-malonyltransferase) coding genes were down regulated, which might explain the increase of aglycones after germination. This study provided a universal transcriptomic atlas for whole soybean seeds germination in terms of nutrition and gene regulation mechanism.

Generalizable Sample-Efficient Siamese Autoencoder for Tinnitus Diagnosis in Listeners With Subjective Tinnitus.

Electroencephalogram (EEG)-based neurofeedback has been widely studied for tinnitus therapy in recent years. Most existing research relies on experts' cognitive prediction, and studies based on machine learning and deep learning are either data-hungry or not well generalizable to new subjects. In this paper, we propose a robust, data-efficient model for distinguishing tinnitus from the healthy state based on EEG-based tinnitus neurofeedback. We propose trend descriptor, a feature extractor with lower fineness, to reduce the effect of electrode noises on EEG signals, and a siamese encoder-decoder network boosted in a supervised manner to learn accurate alignment and to acquire high-quality transferable mappings across subjects and EEG signal channels. Our experiments show the proposed method significantly outperforms state-of-the-art algorithms when analyzing subjects' EEG neurofeedback to 90dB and 100dB sound, achieving an accuracy of 91.67%-94.44% in predicting tinnitus and control subjects in a subject-independent setting. Our ablation studies on mixed subjects and parameters show the method's stability in performance.

C/EBPß regulates the JAK/STAT signaling pathway in triple-negative breast cancer.

C/EBPß is a member of the CCAAT/enhancer-binding protein (C/EBP) family, which consists of a number of b-ZIP transcription factors. Although C/EBPß has been implicated in the development of certain cancers, including breast cancer, it remains unknown whether dysregulation of C/EBPß in breast cancer is subtype-specific. Moreover, the underlying mechanisms by which C/EBPß regulates breast cancer carcinogenesis are not fully understood. Here, we present evidence that C/EBPß is specifically overexpressed in human TNBC samples, but not in non-TNBC samples. C/EBPß depletion dramatically suppressed TNBC cell growth, migration, invasion, and colony formation ability. A subsequent mechanistic study revealed that the JAK/STAT signaling pathway was upregulated in C/EBPß_high TNBC samples compared with C/EBPß_low TNBC samples. C/EBPß ChIP-seq and qPCR were performed to demonstrate that C/EBPß directly binds to and regulates JAK/STAT signaling pathway genes in TNBC. Taken together, our data indicate the oncogenic role of C/EBPß in human TNBC and reveal a novel mechanism by which C/EBPß promotes TNBC carcinogenesis.

[Simulation on soil moisture and water productivity of apple orchard on the Loess Plateau, Northwest China]. / 黄土高原苹果园土壤水分及水分生产力模拟.

The WinEPIC model was used to simulate the dynamics of soil moisture and water productivity in the deep layer of the dry farm apple orchard of Changwu in the Loess Plateau from 1980 to 2018, aiming to provide a scientific basis for the sustainable development of apple production in the area. The results showed that the average annual yield of apple orchards in Changwu area was 27.37 t·hm-2, the average annual evapotranspiration was 673.66 mm, and the average annual water productivity was 4.07 kg·m-3. The number of water stress days in adult apple trees was mainly affected by rainfall. The average number of stress days in the late stage of apple tree growth was 46.46 d. The soil water content in deep layer began to approach withering humidity as early as 9-year-old apple trees. Water supply in the whole growing season of Changwu area was the dominant factor impacting the yield of orchards. The reduction of effective soil water content in deep soil was the main factor restricting yield enhancement in the middle and late growth stages of apple trees. When there was no sufficient precipitation, apple trees would use soil water from deeper soil layer. Excessive precipitation could not be used by apple trees but could be converted into shallow soil moisture and evaporation if the deep layer had less available water. For the mature apple trees, less than 500 mm or higher than 700 mm of annual water supply would cause a decline in production. For apple orchard at different growth periods, water management strategy should be adjusted according to rainfall conditions in different years. Supplementary irrigation, rainwater retention, covering, and pruning of branches could be used to reduce the unproductive and luxury water consumption of apple trees, delay the appearance of deep dry layer of soil, and avoid the waste of water resources while ensuring the growth of apple trees.

Cardiopulmonary haemodynamics after sieve-shaped atrial septal defect repair: a multicenter study.

OBJECTIVE: The study aimed to collect case data on cardiopulmonary bypass (CPB) sieve-shaped (S-S) and non-sieve-shaped (N-S-S) atrial septal defects (ASDs). METHODS: We analysed and summarized the postoperative blood flow in the cardiopulmonary system. We retrospectively collected 86 patients who underwent CPB S-S and N-S-S ASD repair. The data collected included sex, age, CPB time, ASD area, percentage change in ventricular value (PVV) (preoperative/postoperative), left ventricular wall thickness, ejection fraction (EF) (preoperative/postoperative), fluid inflow value, pulmonary arterial pressure/pulmonary venous pressure, percentage change in total lung resistance (PTLR) (preoperative/postoperative) for statistical analysis and comparison. RESULTS: There were 86 eligible patients in this study, 37 and 49 of whom had S-S and N-S-S ASDs, respectively. The PVV, PTLR, and pulmonary arterial pressure/pulmonary venous pressure (postoperative) were significantly different between the S-S and N-S-S groups. The mean PTLR in the S-S and N-S-S groups was 0.78±0.24 and 0.62±0.28, respectively. The mean PVV in the S-S group was 11.53±7.63, and that in the N-S-S group was 16.47±9.71. Multivariate analysis revealed PVV (OR, 0.143; 95% CI, 0.026-0.790; P=0.026), PTLR (OR, 0.156; 95% CI, 0.049-0.500; P=0.002), and pulmonary arterial pressure/pulmonary venous pressure (postoperative) (OR, 9.014; 95% CI, 2.480-32.755; P=0.001) as significant factors. The rate of pulmonary infection absence postoperatively in the S-S group was 76.52%, and that in the N-S-S group was 42.75%. CONCLUSION: Due to the differences in heart structure between the S-S and N-S-S groups, the haemodynamic index (PVV and PTLR, postoperative pulmonary arterial pressure/pulmonary venous pressure) changes after S-S ASD repair were less than those after N-S-S ASD repair, so the postoperative pulmonary infection rate was higher after N-S-S ASD repair. The pulmonary infection rate was low after S-S ASD repair, and drugs should be reasonably administered to prevent infection.

A survey on deep learning-based non-invasive brain signals: recent advances and new frontiers.

Brain signals refer to the biometric information collected from the human brain. The research on brain signals aims to discover the underlying neurological or physical status of the individuals by signal decoding. The emerging deep learning techniques have improved the study of brain signals significantly in recent years. In this work, we first present a taxonomy of non-invasive brain signals and the basics of deep learning algorithms. Then, we provide the frontiers of applying deep learning for non-invasive brain signals analysis, by summarizing a large number of recent publications. Moreover, upon the deep learning-powered brain signal studies, we report the potential real-world applications which benefit not only disabled people but also normal individuals. Finally, we discuss the opening challenges and future directions.

Deep learning for misinformation detection on online social networks: a survey and new perspectives.

Recently, the use of social networks such as Facebook, Twitter, and Sina Weibo has become an inseparable part of our daily lives. It is considered as a convenient platform for users to share personal messages, pictures, and videos. However, while people enjoy social networks, many deceptive activities such as fake news or rumors can mislead users into believing misinformation. Besides, spreading the massive amount of misinformation in social networks has become a global risk. Therefore, misinformation detection (MID) in social networks has gained a great deal of attention and is considered an emerging area of research interest. We find that several studies related to MID have been studied to new research problems and techniques. While important, however, the automated detection of misinformation is difficult to accomplish as it requires the advanced model to understand how related or unrelated the reported information is when compared to real information. The existing studies have mainly focused on three broad categories of misinformation: false information, fake news, and rumor detection. Therefore, related to the previous issues, we present a comprehensive survey of automated misinformation detection on (i) false information, (ii) rumors, (iii) spam, (iv) fake news, and (v) disinformation. We provide a state-of-the-art review on MID where deep learning (DL) is used to automatically process data and create patterns to make decisions not only to extract global features but also to achieve better results. We further show that DL is an effective and scalable technique for the state-of-the-art MID. Finally, we suggest several open issues that currently limit real-world implementation and point to future directions along this dimension.

A Semisupervised Recurrent Convolutional Attention Model for Human Activity Recognition.

Recent years have witnessed the success of deep learning methods in human activity recognition (HAR). The longstanding shortage of labeled activity data inherently calls for a plethora of semisupervised learning methods, and one of the most challenging and common issues with semisupervised learning is the imbalanced distribution of labeled data over classes. Although the problem has long existed in broad real-world HAR applications, it is rarely explored in the literature. In this paper, we propose a semisupervised deep model for imbalanced activity recognition from multimodal wearable sensory data. We aim to address not only the challenges of multimodal sensor data (e.g., interperson variability and interclass similarity) but also the limited labeled data and class-imbalance issues simultaneously. In particular, we propose a pattern-balanced semisupervised framework to extract and preserve diverse latent patterns of activities. Furthermore, we exploit the independence of multi-modalities of sensory data and attentively identify salient regions that are indicative of human activities from inputs by our recurrent convolutional attention networks. Our experimental results demonstrate that the proposed model achieves a competitive performance compared to a multitude of state-of-the-art methods, both semisupervised and supervised ones, with 10% labeled training data. The results also show the robustness of our method over imbalanced, small training data sets.

The Optimized Multi-Scale Permutation Entropy and Its Application in Compound Fault Diagnosis of Rotating Machinery.

Multi-scale permutation entropy (MPE) is a statistic indicator to detect nonlinear dynamic changes in time series, which has merits of high calculation efficiency, good robust ability, and independence from prior knowledge, etc. However, the performance of MPE is dependent on the parameter selection of embedding dimension and time delay. To complete the automatic parameter selection of MPE, a novel parameter optimization strategy of MPE is proposed, namely optimized multi-scale permutation entropy (OMPE). In the OMPE method, an improved Cao method is proposed to adaptively select the embedding dimension. Meanwhile, the time delay is determined based on mutual information. To verify the effectiveness of OMPE method, a simulated signal and two experimental signals are used for validation. Results demonstrate that the proposed OMPE method has a better feature extraction ability comparing with existing MPE methods.