Towards Federated COVID-19 Vaccine Side Effect Prediction

We propose FedCovid, a new federated learning system based on electronic health records (EHR), to predict COVID-19 vaccination side effects. Federated learning allows diverse data owners to work together to train machine learning models without sharing data, ensuring the privacy of EHR data. However, because EHR data is unique, directly using existing federated learning models may fail. The EHR data is diverse, with numerical and categorical characteristics as well as consecutive visits. Furthermore, each client's data size is unequal, and the data labels are skewed due to the small number of patients that experience serious side effects. We present an adaptive approach to fuse heterogeneous EHR data and apply data augmentation techniques working with a margin loss to overcome the data imbalance issue in the client model training to address both challenges simultaneously in FedCovid. We recommend that when the server is updated, the data size of each client be taken into account to lessen the impact of clients with small data volumes. Finally, in order to train a stable and successful federated learning model, we suggest a new ordinal training technique. Experiments on a real-world dataset reveal that the suggested model is effective at predicting COVID-19 vaccination adverse effects. The performance increases by 14.35%, 17.81%, and 129.36% on the F1 score, Cohen's Kappa, and PR-AUC, respectively, compared with the best baseline (The source code of the proposed FedCovid is available at https://github.com/JackqqWang/FedCovid.git ). © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Remote Laboratory as an Educational Tool in Robotics Experimental Course

A remote lab is a technology that allows participants to efficiently conduct experimental teaching where users can connect to lab equipment from anywhere without being in a specific physical location. The COVID-19 pandemic affects all areas of human activity. As a result, students did not receive face-to-face instruction, and access to the laboratory was limited or practically impos-sible, and access to laboratory facilities has been limited or nearly impossible. Especially in engineering education, students' practical abilities cannot be devel-oped comprehensively. In this paper, this paper built an online remote robotics experiment system using digital twin (DT) technology and IoT technology and adopted ADDIE (Analysis, Design, Development, Implementation, and Evalua-tion) teaching method. With these measures, students can design and debug robot programs at home, just like in the laboratory. This study sent questionnaires to 64 students, and 58 were returned. The results show that more than 80% of students believe that the remote labs for industrial robotics courses have improved the efficiency and quality of students' skills training as opposed to virtual simulation and watching videos on the computer.

Shall We Keep Using Zoom etc. after the Pandemic?

Online learning has been studied long before the pandemic. Some educators were concerned about online learning. The COVID-19 pandemic changed everything. During the earlier part of the pandemic (Spring 2020), many universities were locked down. Every course had to be switched to online delivery mode. This imposed quite a challenge for some courses. For example, how would you conduct labs which needed to use lab equipment? How would the students work together on their capstone projects? The successes and lessons learned during the pandemic are an important part of the effort to take advantage of online learning. Many of the practices forced upon us during the pandemic are useful even after the pandemic. Online presentation, use of GitHub for software development, use of Google documents/directory, Google form for team evaluation and peer evaluation are a few things that can be adopted after pandemic to improve student learning. In this paper, successes and lessons learned will be shared regarding the use of Zoom in lectures, laboratories, and help sessions, homework and quizzes in Canvas, virtual presentation for Mini-Maker Faire, feedback from students, and capstone projects. © American Society for Engineering Education, 2022.

Histology-proven Post-COVID Lung Fibrosis

Background: Pulmonary fibrosis is a serious complication of viral pneumonia caused by severe acute respiratory syndrome coronavirus 2 (SARS-COV2). COVID-19 is believed to trigger substantial fibrotic consequences during acute infection. However, the extent to which lung fibrotic change could last and the degree of lung fibrosis in patients with complete resolution of infection still remain ambiguous. A predominant majority of reports on post-COVID-19 pulmonary fibrosis were drawn from radio imaging studies. By contrast, histological evidences of post-COVID-19 pulmonary fibrosis are paradoxically in significant shortage whilst they have higher diagnostic values. We herein report postmortem autopsies focusing on lungs from six patients with resolved SARS-CoV-2 infection. Design: Eligible autopsy samples were collected from patients who died from diseases other than acute COVID-19 and who contracted SARS-CoV-2 virus but either had subsequent negative SARS-CoV-2 test (n=4) or the symptoms of COVID-19 no longer existed and died after at least 100 days after initial positive SARS-CoV-2 test (n=2). Results: These patients included 4 men and 2 women with a mean age of 63 years (range 28 - 79 years). Two patients died from cardiovascular compromise, one patient died from venous thrombosis, one patient died from acute pneumonia, one patient died from post-COVID lung fibrosis, and one patient died from metastatic prostatic adenocarcinoma. Histology of lungs from all six cases showed different degree of fibrosis (Table 1). Remarkably, three of six cases showed extensive patchy interstitial fibrosis. Three of four cases with imaging data reviewed revealed consistent findings in CXR or CT (Table 1). Case (79 yo male) who died from post-COVID lung complications at 410 days after initial positive SARS-COV2 test showed remarkably diffuse lung parenchyma damage with extensive fibrotic changes, honey combing, with an interstitial pattern (Figure 1). Conclusions: Post-COVID fibrotic lung change is present in some patients following resolution of COVID infection. The extent to which lung fibrotic change could last and the degree of lung fibrosis in patients with complete resolution of infection vary from case to case. However, the finding of significant histologically-proven fibrotic lung changes more than 400 days after the resolution of acute COVID-19 in the setting of autopsy provide insight into the pathogenesis and prognosis of long-lasting complications of COVID-19.

Markowitz and Index Models for Optimizing Investment Portfolio Study during COVID-19 Pandemic

Due to the spread of COVID-19 which adversely affects the global economy, the financial market experiences huge changes. To help the investors have a better understanding of the risk in the market before investing, the paper uses the data for 10 stocks from 2001 to 2021 to find out the COVID-19 pandemic's influence on the investment portfolio We construct the Markowitz and Index models under 5 constraints, and then compare the results to illustrate the optimal decisions for investing. The Markowitz model considers expected return and risk, while not attaching enough importance to risk-free assets. The Index model is brief, but the assumption it uses does not have a major influence on the results obtained. Comparing the results, it is obvious that the return and standard deviation of Index model portfolios are higher than the Markowitz model's one. Moreover, the Markowitz performs better under most of the constraints. © 2022 ACM.

Hierarchical Gaussian Processes and Mixtures of Experts to Model COVID-19 Patient Trajectories

Gaussian processes (GPs) are a versatile nonparametric model for nonlinear regression and have been widely used to study spatiotemporal phenomena. However, standard GPs offer limited interpretability and generalizability for datasets with naturally occurring hierarchies. With large-scale, rapidly-updating electronic health record (EHR) data, we want to study patient trajectories across diverse patient cohorts while preserving patient subgroup structure. In this work, we partition our cohort of over 2000 COVID-19 patients by sex and ethnicity. We develop and apply a hierarchical Gaussian process and a mixture of experts (MOE) hierarchical GP model to fit patient trajectories on clinical markers of disease progression. A case study for albumin, an effective predictor of COVID-19 patient outcomes, highlights the predictive performance of these models. These hierarchical spatiotemporal models of EHR data bring us a step closer toward our goal of building flexible approaches to capture patient data that can be used in real-time systems*.

[Evaluation of fast nucleic acid detection system in severe acute respiratory syndrome coronavirus 2].

Objective: To evaluate the performance and application of a fast nucleic acid detection system for testing severe acute respiratory syndrome virus 2 (SARS-COV-2). Methods: Clinical samples were collected from February to July 2020 from Beijing Center for Diseases Prevention and Control and the Laboratory Department of China-Japan Friendship Hospital, to evaluate the sensitivity, specificity, anti-interference ability, precision and clinical sample coincidence rate of fast nucleic acid detection system for SARS-CoV-2. The analytical sensitivity was determined by a dilution series of 20 replications for each concentration. Analytical specificity study was performed by testing organisms whose infection produces symptoms similar to those observed at the onset of corona virus disease 2019 (COVID-19), and of the normal or pathogenic microflora that may be present in specimens collected. Potential interference substances were evaluated with different concentration in the interference study. Precision study was conducted by estimating intra-and inter-batch variability. Clinical evaluation was performed by testing 230 oropharyngeal swab specimens and 95 sputum specimens in fast nucleic acid detection system, comparing with conventional real-time fluorescent quantitative PCR (RT-qPCR) and clinical diagnostic results. Results: The analytical sensitivity of SARS-CoV-2 using fast nucleic acid detection system was 400 copies/ml. The result is negative for testing with the organisms that may likely in the circulating area or causing similar symptoms with SARS-CoV-2 and human nucleic acid, indicating that no cross reactivity with organisms. The results of precision test showed that the Coefficient of variation of Ct value of high, medium and low concentration samples was 1.90%-3.92%, and all of them were less than 5% in intra-and inter-batch testing. The results of the samples were still positive after adding the potential interfering substances, indicating that the possible interfering substances in the samples had no effect on the results. 98.46% and 97.85% diagnosis results of fast nucleic acid detection system were consistent with RT-qPCR and clinical diagnostic results, respectively. Conclusion: The fast nucleic acid detection system based on molecular parallel reaction can be used as a selection method for SARS-CoV-2 testing.

Revealing the potential pharmacological mechanism of traditional Chinese medicine by integrating metabolite profiling of a Q-marker and network pharmacology, prim-O-glucosylcimifugin as an example

Along with definite clinical effects, traditional Chinese medicine (TCM) has increasingly gained worldwide attention. Quality control marker (Q-marker) is always used to confirm the primary effects of herbs or TCM. Thus, the characterization of the metabolism features and subsequently induced effects are essential to reveal the functional consequences. In this work, a strategy by integrating metabolite profiling of a Q-marker and network pharmacology was proposed and applied to explore the potential function of herbs, and prim-O-glucosylcimifugin, the Q-marker of Radix Saposhnikoviae, was used as an example. As a result, 35 metabolites were characterized in rats' plasma, urine, feces, heart, liver, spleen, lung, kidney, and brain. Among them, M13/M17/M20/M23/M24/M28 were the primary metabolites, and 23 metabolites were reported for the first time. It was found that major metabolic pathways of prim-O-glucosylcimifugin were hydroxylation, dehydrogenation, hydrogenation and glucuronidation. In addition to the above metabolic reactions, methylation, hydrogenation, glycosylation, isomerization, sulfation, and other S-conjunctions were found in prim-O-glucosylcimifugin for the first time. Meanwhile, targets of prim-O-glucosylcimifugin and its major metabolites targeted additional 125 targets with functions of MAPK signaling pathway, proteoglycans in cancer, pathways in cancer, bladder cancer, colorectal cancer, serotonergic synapse and natural killer cell-mediated cytotoxicity. Among them, the depression of inflammation by MAPK signaling pathway might be the core mechanism for prim-O-glucosylcimifugin to treat respiratory tract infections. The above results provided vital information for understanding the metabolism and functional mechanism of prim-O-glucosylcimifugin in the treatment of respiratory tract infections, and a new insight for revealing the pharmacological mechanism of the complex system was also provided.

One-Year Real-Time Measurement of Black Carbon in the Rural Area of Qingdao, Northeastern China: Seasonal Variations, Meteorological Effects, and the COVID-19 Case Analysis

In this paper, we report the results obtained from one year of real-time measurement (i.e., from December 2019 to November 2020) of atmospheric black carbon (BC) under a rural environment in Qingdao of Northeastern China. The annual average concentration of BC was 1.92 +/- 1.89 mu g m(-3). The highest average concentration of BC was observed in winter (3.65 +/- 2.66 mu g m(-3)), followed by fall (1.73 +/- 1.33 mu g m(-3)), spring (1.53 +/- 1.33 mu g m(-3)), and summer (0.83 +/- 0.56 mu g m(-3)). A clear weekend effect was observed in winter, which was characterized by higher BC concentration (4.60 +/- 2.86 mu g m(-3)) during the weekend rather than that (3.22 +/- 2.45 mu g m(-3)) during weekdays. The influence of meteorological parameters, including surface horizontal wind speed, boundary layer height (BLH), and precipitation, on BC, was investigated. In particular, such BLH influence presented evidently seasonal dependence, while there was no significant seasonality for horizontal wind speed. These may reflect different roles of atmospheric vertical dilution on affecting BC in different seasons. The oBC/oCO ratio decreased with the increase of precipitation, indicative of the influence of below-cloud wet removal of BC, especially during summertime where rainfall events more frequently occurred than any of other seasons. The bivariate-polar-plot analysis showed that the high BC concentrations were mainly associated with low wind speed in all seasons, highlighting an important BC source originated from local emissions. By using concentration-weighted trajectory analysis, it was found that regional transports, especially from northeastern in winter, could not be negligible for contributing to BC pollution in rural Qingdao. In the coronavirus disease 2019 (COVID-19) case analysis, we observed an obvious increase in the BC/NO2 ratio during the COVID-19 lockdown, supporting the significant non-traffic source sector (such as residential coal combustion) for BC in rural Qingdao.

[Epidemiological characteristics of COVID-19 cases in outbreak in Xinfadi market in Beijing].

Objective: To understand the epidemiological characteristics of COVID-19 cases, including asymptomatic cases and symptomatic cases, in the outbreak in Xinfadi market in Beijing. Methods: Data and epidemiological survey reports of COVID-19 cases in Xinfadi market in Beijing were extracted from China's Infectious Disease Information System. Epidemiological characteristics of symptomatic cases and asymptomatic cases were analyzed and compared by using software SPSS 19.0. Results: From June 11 to July 10, 2020, a total of 368 laboratory-confirmed COVID-19 cases reported in Xinfadi market, in which, 335 (91.03%) were symptomatic and 33 (8.97%) were asymptomatic. The cases were distributed in 11 districts, and most cases (252/368, 68.48%) were reported in Fengtai district. The incidence curve of the cases showed a typical outbreak pattern, the case number peaked on 13 June. The median age of the cases were 43 years (QR: 31-51). The asymptomatic cases (M=32, QR: 29-46) were younger than the symptomatic cases (M=43, QR: 31-52), the difference was significant (Z=2.416, P=0.016). The ratio of male to female was 1.26∶1. Most cases (236/368, 64.13%) were engaged in catering service and public place service. 73.91% (272/368) of the cases had historg of direct exposures in the Xinfadi market. 54.08%(199/368) of cases were detected through nucleic acid testing and screening. Mild and moderate cases accounted for 99.10% (332/335) and no death occurred. Conclusion: The COVID-19 cases in the outbreak in Xinfadi market were mainly engaged in catering service and public place service. The asymptomatic cases were younger than the symptomatic cases.