ConTIG: Continuous representation learning on temporal interaction graphs.

Representation learning on temporal interaction graphs (TIG) aims to model complex networks with the dynamic evolution of interactions on a wide range of web and social graph applications. However, most existing works on TIG either (a) rely on discretely updated node embeddings merely when an interaction occurs that fail to capture the continuous evolution of embedding trajectories of nodes, or (b) overlook the rich temporal patterns hidden in the ever-changing graph data that presumably lead to sub-optimal models. In this paper, we propose a two-module framework named ConTIG, a novel representation learning method on TIG that captures the continuous dynamic evolution of node embedding trajectories. With two essential modules, our model exploits three-fold factors in dynamic networks including latest interaction, neighbor features, and inherent characteristics. In the first update module, we employ a continuous inference block to learn the nodes' state trajectories from time-adjacent interaction patterns using ordinary differential equations. In the second transform module, we introduce a self-attention mechanism to predict future node embeddings by aggregating historical temporal interaction information. Experiment results demonstrate the superiority of ConTIG on temporal link prediction, temporal node recommendation, and dynamic node classification tasks of four datasets compared with a range of state-of-the-art baselines, especially for long-interval interaction prediction.

Advancing safe mobility: A global analysis of research trends in safe route planning.

Safe route planning has become an increasingly important area of research in recent years due to growing concerns about pedestrian and traffic safety, rising traffic volumes and densities in urban areas, and advancements in smart vehicle and transportation technologies. This study conducted a bibliometric analysis of publications on safe route planning retrieved from the Web of Science database between January 2000 and January 2023 to understand the state of the field. A total of 1546 publications authored by 5423 researchers from 84 countries were analyzed. The findings identified the United States, China, India, South Korea, and Spain as the most productive countries, while the University of North Carolina emerged as the most productive organization. Engineering, computer science, transportation, public health, and automation were revealed to be the dominant initial research areas, although interest grew from other domains like urban planning and the environment over time. Analysis of publications by year showed a steady rise in output starting from 2008. Notable influential publications and highly cited authors in the field were also identified. Several research themes and terms like path planning, safety, walking, and route to school were highlighted through keyword analysis. This study provided novel insights into the evolving international landscape, topics, and influential contributors in safe route planning research over the past two decades. Limitations in database coverage and analytical techniques necessitate future work to enhance understanding in this critical domain.

A Machine Learning-Based Risk Prediction Model for Post-Traumatic Stress Disorder during the COVID-19 Pandemic.

Background and Objectives: The COVID-19 pandemic has caused global public panic, leading to severe mental illnesses, such as post-traumatic stress disorder (PTSD). This study aimed to establish a risk prediction model of PTSD based on a machine learning algorithm to provide a basis for the extensive assessment and prediction of the PTSD risk status in adults during a pandemic. Materials and Methods: Model indexes were screened based on the cognitive-phenomenological-transactional (CPT) theoretical model. During the study period (1 March to 15 March 2020), 2067 Chinese residents were recruited using Research Electronic Data Capture (REDCap). Socio-demographic characteristics, PTSD, depression, anxiety, social support, general self-efficacy, coping style, and other indicators were collected in order to establish a neural network model to predict and evaluate the risk of PTSD. Results: The research findings showed that 368 of the 2067 participants (17.8%) developed PTSD. The model correctly predicted 90.0% (262) of the outcomes. Receiver operating characteristic (ROC) curves and their associated area under the ROC curve (AUC) values suggested that the prediction model possessed an accurate discrimination ability. In addition, depression, anxiety, age, coping style, whether the participants had seen a doctor during the COVID-19 quarantine period, and self-efficacy were important indexes. Conclusions: The high prediction accuracy of the model, constructed based on a machine learning algorithm, indicates its applicability in screening the public mental health status during the COVID-19 pandemic quickly and effectively. This model could also predict and identify high-risk groups early to prevent the worsening of PTSD symptoms.

Open-top light-sheet imaging of CLEAR emulsion for high-throughput loss-free analysis of massive fluorescent droplets.

Digital droplet PCR (ddPCR) is classified as the third-generation PCR technology that enables absolute quantitative detection of nucleic acid molecules and has become an increasingly powerful tool for clinic diagnosis. We previously established a CLEAR-dPCR technique based on the combination of CLEAR droplets generated by micro-centrifuge-based microtubule arrays (MiCA) andinsitu3D readout by light-sheet fluorescence imaging. This CLEAR-dPCR technique attains very high readout speed and dynamic range. Meanwhile, it is free from sample loss and contamination, showing its advantages over commercial d-PCR technologies. However, a conventional orthogonal light-sheet imaging setup in CLEAR d-PCR cannot image multiple centrifuge tubes, thereby limiting its widespread application to large-scale, high-speed dd-PCR assays. Herein, we propose an in-parallel 3D dd-PCR readout technique based on an open-top light-sheet microscopy setup. This approach can continuously scan multiple centrifuge tubes which contain CLEAR emulsions with highly diverse concentrations, and thus further boost the scale and throughput of our 3D dd-PCR technique.

Isotropic super-resolution light-sheet microscopy of dynamic intracellular structures at subsecond timescales.

Long-term visualization of the dynamic interactions between intracellular structures throughout the three-dimensional space of whole live cells is essential to better understand their functions, but this task remains challenging due to the limitations of existing three-dimensional fluorescence microscopy techniques, such as an insufficient axial resolution, low volumetric imaging rate and photobleaching. Here, we present the combination of a progressive deep-learning super-resolution strategy with a double-ring-modulated selective plane illumination microscopy design capable of visualizing the dynamics of intracellular structures in live cells for hours at an isotropic spatial resolution of roughly 100 nm in three dimensions at speeds up to roughly 17 Hz. Using this approach, we reveal the complex spatial relationships and interactions between endoplasmic reticulum (ER) and mitochondria throughout live cells, providing new insights into ER-mediated mitochondrial division. We also examined the motion of Drp1 oligomers involved in mitochondrial fission and revealed the dynamic interactions between Drp1 and mitochondria in three dimensions.

Complete genome sequence of a porcine parvovirus strain isolated in central china.

We report here the complete genome sequence of the porcine parvovirus (PPV) strain J-PPV, isolated from central China. Our data, together with sequence data for PPV isolates from other regions of China, will help in understanding the epidemiology and molecular characteristics of PPV field isolates in China.