Clinical characteristics of 4,520 paediatric patients infected with the SARS-CoV-2 omicron variant, in Xi'an, China.

Background and objective: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has broad tissue tropism and high transmission, which are likely to perpetuate the pandemic. The study aim to analyze the clinicopathogenic characteristics in paediatric patients. Methods: In this single-centre study, we retrospectively included all confirmed cases infected by SARS-CoV-2 infection at Xi'an Children's Hospital, China, from 1 December to 31 December 2022. The demographic, clinical, laboratory, and radiological features of the patients were analysed. Results: A total of 4,520 paediatric patients with SARS-CoV-2 omicron variant infections were included. Of these, 3,861 (85.36%) were outpatients, 659 (14.64%) were hospitalised patients, and nine patients (0.20%) died. Of the nine patients who died, five were diagnosed with acute necrotising encephalopathy (ANE). The most common symptoms were fever in 4,275 (94.59%) patients, cough in 1,320 (29.20%) patients, convulsions in 610 (13.50%) patients, vomiting in 410 (9.07%) patients, runny nose/coryza in 277 (6.13%) patients, hoarseness of voice in 273 (6.04%) patients. A blood cell analysis showed a slight elevation of monocytes (mean: 11.14 ± 0.07%). The main diagnoses for both outpatients and inpatients were respiratory infection with multisystem manifestations. Conclusions: A high incidence of convulsions is a typical characteristic of children infected with SARS-CoV-2. Five of the nine COVID-19 fatalities were associated with ANE. This indicates that nervous system damage in children with SARS-CoV-2 infection is more significant.

A regression model combined convolutional neural network and recurrent neural network for electroencephalogram-based cross-subject fatigue detection.

Fatigue, one of the most important factors affecting road safety, has attracted many researchers' attention. Most existing fatigue detection methods are based on feature engineering and classification models. The feature engineering is greatly influenced by researchers' domain knowledge, which will lead to a poor performance in fatigue detection, especially in cross-subject experiment design. In addition, fatigue detection is often simplified as a classification problem of several discrete states. Models based on deep learning can realize automatic feature extraction without the limitation of researcher's domain knowledge. Therefore, this paper proposes a regression model combined convolutional neural network and recurrent neural network for electroencephalogram-based (EEG-based) cross-subject fatigue detection. At the same time, a twofold random-offset zero-overlapping sampling method is proposed to train a bigger model and reduce overfitting. Compared with existing results, the proposed method achieves a much better result of 0.94 correlation coefficient (COR) and 0.09 root mean square error (RMSE) in a within-subject experiment design. What is more, there is no misclassification between awake and drowsy states. For cross-subject experiment design, the COR and RMSE are 0.79 and 0.15, respectively, which are close to the existing within-subject results and better than similar cross-subject results. The cross-subject regression model is very important for fatigue detection application since the fatigue indication is more precise than several discrete states and no model calibration is required for a new user. The twofold random-offset zero-overlapping sampling method can also be used as a reference by other EEG-based deep learning research.

A regression method for EEG-based cross-dataset fatigue detection.

Introduction: Fatigue is dangerous for certain jobs requiring continuous concentration. When faced with new datasets, the existing fatigue detection model needs a large amount of electroencephalogram (EEG) data for training, which is resource-consuming and impractical. Although the cross-dataset fatigue detection model does not need to be retrained, no one has studied this problem previously. Therefore, this study will focus on the design of the cross-dataset fatigue detection model. Methods: This study proposes a regression method for EEG-based cross-dataset fatigue detection. This method is similar to self-supervised learning and can be divided into two steps: pre-training and the domain-specific adaptive step. To extract specific features for different datasets, a pretext task is proposed to distinguish data on different datasets in the pre-training step. Then, in the domain-specific adaptation stage, these specific features are projected into a shared subspace. Moreover, the maximum mean discrepancy (MMD) is exploited to continuously narrow the differences in the subspace so that an inherent connection can be built between datasets. In addition, the attention mechanism is introduced to extract continuous information on spatial features, and the gated recurrent unit (GRU) is used to capture time series information. Results: The accuracy and root mean square error (RMSE) achieved by the proposed method are 59.10% and 0.27, respectively, which significantly outperforms state-of-the-art domain adaptation methods. Discussion: In addition, this study discusses the effect of labeled samples. When the number of labeled samples is 10% of the total number, the accuracy of the proposed model can reach 66.21%. This study fills a vacancy in the field of fatigue detection. In addition, the EEG-based cross-dataset fatigue detection method can be used for reference by other EEG-based deep learning research practices.

The Landscape of Accessible Chromatin and Developmental Transcriptome Maps Reveal a Genetic Mechanism of Skeletal Muscle Development in Pigs.

The epigenetic regulation mechanism of porcine skeletal muscle development relies on the openness of chromatin and is also precisely regulated by transcriptional machinery. However, fewer studies have exploited the temporal changes in gene expression and the landscape of accessible chromatin to reveal the underlying molecular mechanisms controlling muscle development. To address this, skeletal muscle biopsy samples were taken from Landrace pigs at days 0 (D0), 60 (D60), 120 (D120), and 180 (D180) after birth and were then analyzed using RNA-seq and ATAC-seq. The RNA-seq analysis identified 8554 effective differential genes, among which ACBD7, TMEM220, and ATP1A2 were identified as key genes related to the development of porcine skeletal muscle. Some potential cis-regulatory elements identified by ATAC-seq analysis contain binding sites for many transcription factors, including SP1 and EGR1, which are also the predicted transcription factors regulating the expression of ACBD7 genes. Moreover, the omics analyses revealed regulatory regions that become ectopically active after birth during porcine skeletal muscle development after birth and identified 151,245, 53,435, 30,494, and 40,911 peaks. The enriched functional elements are related to the cell cycle, muscle development, and lipid metabolism. In summary, comprehensive high-resolution gene expression maps were developed for the transcriptome and accessible chromatin during postnatal skeletal muscle development in pigs.

Research on improved intestinal image classification for LARS based on ResNet.

Low anterior rectal resection is an effective way to treat rectal cancer at present, but it is easy to cause low anterior resection syndrome after surgery; so, a comprehensive diagnosis of defecation and pelvic floor function must be carried out. There are few studies on the classification of diagnoses in the field of intestinal diseases. In response to these outstanding problems, this research will focus on the design of the intestinal function diagnosis system and the image processing and classification algorithm of the intestinal wall to verify an efficient fusion method, which can be used to diagnose the intestinal diseases in clinical medicine. The diagnostic system designed in this paper makes up for the singleness of clinical monitoring methods. At the same time, the Res-SVDNet neural network model is used to solve the problems of small intestinal image samples and network overfitting, and achieve efficient fusion diagnosis of intestinal diseases in patients. Different models were used to compare experiments on the constructed datasets to verify the applicability of the Res-SVDNet model in intestinal image classification. The accuracy of the model was 99.54%, which is several percentage points higher than other algorithm models.

Calibration of camera internal parameters based on grey wolf optimization improved by levy flight and mutation.

Traditional calibration technology has been widely used in measurement and monitoring; however, there are limitations of poor calibration accuracy, which can not meet the accuracy requirements in some scenarios. About this problem, we proposed a grey wolf optimization algorithm based on levy flight and mutation mechanism to solve camera internal parameters in this paper. The algorithm is based on the actual nonlinear model, which takes the minimum average value of reprojection error as the objective function. The grey wolf position is randomly generated within a given range. Then, the grey wolf optimization algorithm based on levy flight and mutation mechanism is used to iteratively calculate the optimal position, which is the internal parameters of cameras. The two groups of experimental data were performed to verify the algorithm. The result shows better effectiveness and calibration accuracy of the proposed algorithm compared with other optimization methods.

The landscape of chromatin accessibility in skeletal muscle during embryonic development in pigs.

BACKGROUND: The development of skeletal muscle in pigs during the embryonic stage is precisely regulated by transcriptional mechanisms, which depend on chromatin accessibility. However, how chromatin accessibility plays a regulatory role during embryonic skeletal muscle development in pigs has not been reported. To gain insight into the landscape of chromatin accessibility and the associated genome-wide transcriptome during embryonic muscle development, we performed ATAC-seq and RNA-seq analyses of skeletal muscle from pig embryos at 45, 70 and 100 days post coitus (dpc). RESULTS: In total, 21,638, 35,447 and 60,181 unique regions (or peaks) were found across the embryos at 45 dpc (LW45), 70 dpc (LW70) and 100 dpc (LW100), respectively. More than 91% of the peaks were annotated within - 1 kb to 100 bp of transcription start sites (TSSs). First, widespread increases in specific accessible chromatin regions (ACRs) from embryos at 45 to 100 dpc suggested that the regulatory mechanisms became increasingly complicated during embryonic development. Second, the findings from integrated ATAC-seq and RNA-seq analyses showed that not only the numbers but also the intensities of ACRs could control the expression of associated genes. Moreover, the motif screening of stage-specific ACRs revealed some transcription factors that regulate muscle development-related genes, such as MyoG, Mef2c, and Mef2d. Several potential transcriptional repressors, including E2F6, OTX2 and CTCF, were identified among the genes that exhibited different regulation trends between the ATAC-seq and RNA-seq data. CONCLUSIONS: This work indicates that chromatin accessibility plays an important regulatory role in the embryonic muscle development of pigs and regulates the temporal and spatial expression patterns of key genes in muscle development by influencing the binding of transcription factors. Our results contribute to a better understanding of the regulatory dynamics of genes involved in pig embryonic skeletal muscle development.

Genome-Wide Association Analysis Identified BMPR1A as a Novel Candidate Gene Affecting the Number of Thoracic Vertebrae in a Large White × Minzhu Intercross Pig Population.

The number of vertebrae (NV), especially the number of thoracic vertebrae (NTV), varies among pig breeds. The NTV is controlled by vertebral segmentation and the number of somites during embryonic development. Although there is a high correlation between the NTV and NV, studies on a fixed NV have mainly considered the absolute numbers of thoracic vertebrae instead of vertebral segmentation. Therefore, this study aimed to discover variants associated with the NTV by considering the effect of the NV in pigs. The NTV and NV of 542 F2 individuals from a Large White × Minzhu pig crossbreed were recorded. All animals were genotyped for VRTN g.19034 A > C, LTBP2 c.4481A > C, and 37 missense or splice variants previously reported in a 951-kb interval on SSC7 and 147 single nucleotide polymorphisms (SNPs) on SSC14. To identify NTV-associated SNPs, we firstly performed a genome-wide association study (GWAS) using the Q + K (population structure + kinship matrix) model in TASSEL. With the NV as a covariate, the obtained data were used to identify the SNPs with the most significant genome-wide association with the NTV by performing a GWAS on a PorcineSNP60K Genotyping BeadChip. Finally, a conditional GWAS was performed by fixing this SNP. The GWAS showed that 31 SNPs on SSC7 have significant genome-wide associations with the NTV. No missense or splice variants were found to be associated with the NTV significantly. A linkage disequilibrium analysis suggested the existence of quantitative trait loci (QTL) in a 479-Kb region on SSC7, which contained a critical candidate gene FOS for the NTV in pigs. Subsequently, a conditional GWAS was performed by fixing M1GA0010658, the most significant of these SNPs. Two SNPs in BMPR1A were found to have significant genome-wide associations and a significant dominant effect. The leading SNP, S14_87859370, accounted for 3.86% of the phenotypic variance. Our study uncovered that regulation variants in FOS on SSC7 and in BMPR1A on SSC14 might play important roles in controlling the NTV, and thus these genetic factors may be harnessed for increasing the NTV in pigs.

Transcriptome Analysis of Skeletal Muscle in Pigs with Divergent Residual Feed Intake Phenotypes.

Residual feed intake (RFI) is defined as the difference between the observed and expected feed intake for maintenance and growth requirements. In this study, the expression profiles of mRNAs and long noncoding RNAs (lncRNAs) from skeletal muscle in Duroc pigs with divergent RFI phenotypes were investigated by Illumina sequencing. Finally, a total of 2195 annotated lncRNAs and 1976 novel lncRNAs were obtained. About 210 mRNAs and 43 lncRNAs were differentially expressed among high and low RFI pigs. The differentially expressed mRNAs were potentially involved in the biological processes of lipid metabolism, extracellular matrix organization, cell proliferation, and cell adhesion. The lipolysis in skeletal muscle was increased in high RFI pigs, suggesting that high RFI pigs might need more energy than low RFI pigs. However, skeletal muscle development was increased in low RFI pigs. These results suggested that low RFI pigs might be more efficient in energy utilization during skeletal muscle growth. The function of lncRNA was also analyzed by target prediction. Nine lncRNAs might be candidate lncRNAs for the determination of RFI phenotype, by the regulation of the biological processes of lipid metabolism, cell proliferation, and cell adhesion. This study should facilitate a further understanding of the molecular mechanism for the determination of RFI phenotype in pigs.

Genome-wide identification of RNA editing in seven porcine tissues by matched DNA and RNA high-throughput sequencing.

BACKGROUND: RNA editing is a co/posttranscriptional modification mechanism that increases the diversity of transcripts, with potential functional consequences. The advent of next-generation sequencing technologies has enabled the identification of RNA edits at unprecedented throughput and resolution. However, our knowledge of RNA editing in swine is still limited. RESULTS: Here, we utilized RES-Scanner to identify RNA editing sites in the brain, subcutaneous fat, heart, liver, muscle, lung and ovary in three 180-day-old Large White gilts based on matched strand-specific RNA sequencing and whole-genome resequencing datasets. In total, we identified 74863 editing sites, and 92.1% of these sites caused adenosine-to-guanosine (A-to-G) conversion. Most A-to-G sites were located in noncoding regions and generally had low editing levels. In total, 151 A-to-G sites were detected in coding regions (CDS), including 94 sites that could lead to nonsynonymous amino acid changes. We provide further evidence supporting a previous observation that pig transcriptomes are highly editable at PRE-1 elements. The number of A-to-G editing sites ranged from 4155 (muscle) to 25001 (brain) across the seven tissues. The expression levels of the ADAR enzymes could explain some but not all of this variation across tissues. The functional analysis of the genes with tissue-specific editing sites in each tissue revealed that RNA editing might play important roles in tissue function. Specifically, more pathways showed significant enrichment in the fat and liver than in other tissues, while no pathway was enriched in the muscle. CONCLUSIONS: This study identified a total of 74863 nonredundant RNA editing sites in seven tissues and revealed the potential importance of RNA editing in tissue function. Our findings largely extend the porcine editome and enhance our understanding of RNA editing in swine.

Biometric recognition via texture features of eye movement trajectories in a visual searching task.

Biometric recognition technology based on eye-movement dynamics has been in development for more than ten years. Different visual tasks, feature extraction and feature recognition methods are proposed to improve the performance of eye movement biometric system. However, the correct identification and verification rates, especially in long-term experiments, as well as the effects of visual tasks and eye trackers' temporal and spatial resolution are still the foremost considerations in eye movement biometrics. With a focus on these issues, we proposed a new visual searching task for eye movement data collection and a new class of eye movement features for biometric recognition. In order to demonstrate the improvement of this visual searching task being used in eye movement biometrics, three other eye movement feature extraction methods were also tested on our eye movement datasets. Compared with the original results, all three methods yielded better results as expected. In addition, the biometric performance of these four feature extraction methods was also compared using the equal error rate (EER) and Rank-1 identification rate (Rank-1 IR), and the texture features introduced in this paper were ultimately shown to offer some advantages with regard to long-term stability and robustness over time and spatial precision. Finally, the results of different combinations of these methods with a score-level fusion method indicated that multi-biometric methods perform better in most cases.

Mirror-normal difference in the late phase of mental rotation: An ERP study.

Mirror-normal letter discriminations are thought to require mental rotation in order to transform the rotated alphanumeric character into its canonical orientation. Moreover, out-of-plane rotation is likely to occur after in-plane rotation to fully normalize the mirror version before the final mirror-normal judgment. The so-called rotation-related negativity, which varies with orientation, is found in both ERPonset (averaged with respect to stimulus onset) and ERPRT (averaged with respect to response time), representing the involvement of mental rotation in both time windows. Additionally, the mean amplitude of ERPRT correlates with individual performance. We performed a comprehensive analysis of the mirror-normal differences in the early and late phases of mental rotation and deduced that out-of-plane rotation is more likely to occur in the late phase and interacts with both in-plane rotation and the decision-making process, as indicated by both behavioral and electrophysiological findings.

Uncovering the cognitive processes underlying mental rotation: an eye-movement study.

Mental rotation is an important paradigm for spatial ability. Mental-rotation tasks are assumed to involve five or three sequential cognitive-processing states, though this has not been demonstrated experimentally. Here, we investigated how processing states alternate during mental-rotation tasks. Inference was carried out using an advanced statistical modelling and data-driven approach - a discriminative hidden Markov model (dHMM) trained using eye-movement data obtained from an experiment consisting of two different strategies: (I) mentally rotate the right-side figure to be aligned with the left-side figure and (II) mentally rotate the left-side figure to be aligned with the right-side figure. Eye movements were found to contain the necessary information for determining the processing strategy, and the dHMM that best fit our data segmented the mental-rotation process into three hidden states, which we termed encoding and searching, comparison, and searching on one-side pair. Additionally, we applied three classification methods, logistic regression, support vector model and dHMM, of which dHMM predicted the strategies with the highest accuracy (76.8%). Our study did confirm that there are differences in processing states between these two of mental-rotation strategies, and were consistent with the previous suggestion that mental rotation is discrete process that is accomplished in a piecemeal fashion.

A crucial temporal accuracy test of combining EEG and Tobii eye tracker.

Eye tracking and event-related potentials have been widely used in the field of cognitive psychology and neuroscience. Both techniques have the ability to refine cognitive models through a precise timeline description; nevertheless, they also have severe limitations. Combining measures of event-related potentials and eye movements can contribute to cognitive process capture, which provides the possibility to determine precisely when and in which order different cognitive operations occur. Combining of event-related potentials and eye movements has been recently conducted by synchronizing measures from an infrared eye tracker with an electroencephalograph to allow simultaneous data recording. Here, we describe in detail 4 types of co-registration methods for event-related potentials and eye movements on the Tobii platform. Moreover, the present investigation was designed to evaluate the temporal accuracy of data obtained using the 4 methods. We found that the method based on the Tobii Pro Analytics software development kit had a higher degree of temporal accuracy than the other co-registration methods. Furthermore, the reasons for the different temporal accuracies were assessed, and potential measures to correct clock drift were taken. General suggestions are made regarding timing in the co-registration of the electroencephalograph and eye tracker.

Genome-wide study refines the quantitative trait locus for number of ribs in a Large White × Minzhu intercross pig population and reveals a new candidate gene.

In China, sparerib is one of the most valuable parts of the pork carcass. As a result, candidate gene mining for number of ribs has become an interesting study focus. To examine the genetic basis for this major trait, we genotyped 596 individuals from an F2 Large White × Minzhu intercross pig population using the PorcineSNP60 Genotyping BeadChip. The genome-wide association study identified a locus for number of ribs in a 2.38-Mb region on Sus scrofa chromosome 7 (SSC7 of Sus scrofa genome assembly, Sscrofa10.2). We identified the top significant SNP ASGA0035536, which explained 16.51 % of the phenotypic variance. A previously reported candidate causal mutation (g.19034 A>C) in vertebrae development-associated gene VRTN explained 8.79 % of the phenotypic variation on number of ribs and had a much lower effect than ASGA0035536. Haplotype sharing analysis in F1 boars localized the rib number QTL to a 951-kb interval on SSC7. This interval encompassed 17 annotated genes in Sscrofa10.2, including the previously reported VRTN candidate gene. Of the 17 candidate genes, LTBP2, which encodes a latent transforming growth factor beta binding protein, was previously reported to indirectly regulate the activity of growth differentiation factor Gdf11, which has been shown to increase the number of ribs in knock-out mice. Thus, we propose LTBP2 as a good new candidate gene for number of ribs in the pig population. This finding advances our understanding of the genetic architecture of rib number in pigs.

Balancing a simulated inverted pendulum through motor imagery: an EEG-based real-time control paradigm.

Most brain-computer interfaces (BCIs) are non-time-restraint systems. However, the method used to design a real-time BCI paradigm for controlling unstable devices is still a challenging problem. This paper presents a real-time feedback BCI paradigm for controlling an inverted pendulum on a cart (IPC). In this paradigm, sensorimotor rhythms (SMRs) were recorded using 15 active electrodes placed on the surface of the subject's scalp. Subsequently, common spatial pattern (CSP) was used as the basic filter to extract spatial patterns. Finally, linear discriminant analysis (LDA) was used to translate the patterns into control commands that could stabilize the simulated inverted pendulum. Offline trainings were employed to teach the subjects to execute corresponding mental tasks, such as left/right hand motor imagery. Five subjects could successfully balance the online inverted pendulum for more than 35s. The results demonstrated that BCIs are able to control nonlinear unstable devices. Furthermore, the demonstration and extension of real-time continuous control might be useful for the real-life application and generalization of BCI.