Network Public Opinion Detection During the Coronavirus Pandemic: A Short-Text Relational Topic Model

Online social media provides rich and varied information reflecting the significant concerns of the public during the coronavirus pandemic. Analyzing what the public is concerned with from social media information can support policy-makers to maintain the stability of the social economy and life of the society. In this article, we focus on the detection of the network public opinions during the coronavirus pandemic. We propose a novel Relational Topic Model for Short texts (RTMS) to draw opinion topics from social media data. RTMS exploits the feature of texts in online social media and the opinion propagation patterns among individuals. Moreover, a dynamic version of RTMS (DRTMS) is proposed to capture the evolution of public opinions. Our experiment is conducted on a real-world dataset which includes 67,592 comments from 14,992 users. The results demonstrate that, compared with the benchmark methods, the proposed RTMS and DRTMS models can detect meaningful public opinions by leveraging the feature of social media data. It can also effectively capture the evolution of public concerns during different phases of the coronavirus pandemic. © 2021 Association for Computing Machinery.

Effects of SARS-CoV-2 Membrane Glycoprotein on Host Pre-mRNA 3' Processing

Objective To investigate the effect of SARS-CoV-2 membrane protein on the processing of the 3' untranslated region (UTR) of the mRNA precursor (pre-mRNA) in host cells. Methods Based on the cell model of human lung epithelial cells A549, over-expression of the SARS-CoV-2 membrane protein was performed. The RNA-Seq high-throughput sequencing technique and bioinformatics methods was employed to analyze the systematic characterization of alternative polyadenylation (APA) events in host cells. Genes with significant APA events were uploaded to the Metascape database for functional enrichment analysis. In addition, alternative 3'UTR length of genes with APA events was verified by RT-qPCR. Then the target protein expression level was detected by Western blot. Results A total of 813 genes that were significant dynamic APA events in host cells that over-expressed SARS-CoV-2 membrane protein. These genes were enriched in cell biologicial processes such as the mitotic cell cycle and regulation of cellular response to stress. We further screened AKT1, which encodes a critical regulator involved in the above biological process, showing a 3'UTR lengthening in IGV software. RT-qPCR verified the trend of 3'UTR length changes of AKT1. Western blot showed the increased level of phosphorylated AKT1 protein in over-expressed group of M protein. Conclusion SARS-CoV-2 membrane protein potentially affects the 3' processing of host pre-mRNAs. AKT1, which is involved in a variety of viral biological processes, with 3'UTR lengthening, and its protein function was activated intracellularly. © 2022 Institute of Biophysics,Chinese Academy of Sciences. All rights reserved.

Pharm Aco Kin Etics, Safety, and Tolerability of Upadacitinib in Children with Atopic Dermatitis

BACKGROUND: Upadacitinib is a Janus kinase inhibitor that has been approved for the treatment of adults and adolescents with moderate to severe atopic dermatitis (AD). The objective of this study was to characterize the pharmacokinetics (PK), safety, and tolerability of upadacitinib in children with severe atopic dermatitis. METHOD(S): This is an open-label, multiple-dose study. AD patients (n = 35) were enrolled into four cohorts (Cohort 1, 6 to <12 years, low dose;Cohort 2, 6 to <12 years, high dose;Cohort 3, 2 to <6 years, low dose;Cohort 4, 2 to <6 years, high dose). The low and high doses were selected based on body weight to provide comparable plasma exposure in pediatrics to 15 mg and 30 mg QD doses in adults, respectively. All patients continued on the low dose after the PK assessment on Study Day 7. Safety and exploratory efficacy parameters are assessed in the study. RESULT(S): Geometric mean Cmax and AUC over 0-24 hours at steady state were 33.1 ng/mL and 249 ng.h/mL, respectively, in Cohort 1, 95.5 ng/mL and 523 ng.h/mL, respectively, in Cohort 2, 35.2 ng/mL and 264 ng.h/mL, respectively, in Cohort 3, and 101 ng/mL and 625 ng.h/mL, respectively, in Cohort 4. Upadacitinib was generally safe and well tolerated. The most common AEs were COVID infection, headache, and abdominal discomfort. No new safety risks were identified compared to the known safety profile for upadacitinib. In the 29 subjects with available interim efficacy results at week 12, 34.5% achieved validated Investigator's Global Assessment scale for AD score of 0 or 1 and 69.0% achieved Eczema Area and Severity Index by at least 75% at Week 12 with treatment of upadacitinib. CONCLUSION(S): The findings supported the use of current dosing regimens for further investigation of upadacitinib in upcoming phase 3 clinical trials in pediatric AD patients.

Integration of a hybrid scan approach and in-house high-resolution MS2 spectral database for charactering the multicomponents of Xuebijing Injection

Xuebijing (XBJ) Injection is a reputable patent Chinese medicine widely used to cure sepsis, among the Chinese ″Three Medicines and Three Prescriptions″ solution to fight against COVID-19. We were aimed to achieve the comprehensive multicomponent characterization from the single drugs to traditional Chinese medicine (TCM) formula, by integrating powerful data acquisition and the in-house MS2 spectral database searching. By ultra-high performance liquid chromatography/ion mobility-quadrupole time-of-flight mass spectrometry (UHPLC/IM-QTOF-MS), a hybrid scan approach (HDMSE-HDDDA) was developed, while the HDMSE data for five component drugs and 56 reference compounds were acquired and processed to establish an in-house MS2 spectral database of XBJ. Good resolution of the XBJ components was accomplished on a Zorbax Eclipse Plus C18 column within 24 min, while a fit-for-purpose HDMSE-HDDDA approach was elaborated in two ionization modes for enhanced MS2 data acquisition. XBJ MS2 spectral library was thus established on the UNIFITM platform involving rich structure-related information for the chemicals from five component drugs. We could identify or tentatively characterize 294 components from XBJ, involving 81 flavonoids, 51 terpenoids, 42 phthalides, 40 organic acids, 13 phenylpropanoids, seven phenanthrenequinones, six alkaloids, and 54 others. In contrast to the application of conventional MS1 library, this newly established strategy could demonstrate superiority in the accuracy of identification results and the characterization of isomers, due to the more restricted filtering/matching criteria. Conclusively, the integration of the HDMSE-HDDDA hybrid scan approach and the in-house MS2 spectral database can favor the efficient and more reliable multicomponent characterization from single drugs to the TCM formula. © 2022 The Author(s)

Identification and development of a five-gene signature to improve the prediction of mechanical ventilator-free days for patients with COVID-19.

OBJECTIVE: Coronavirus disease 2019 (COVID-19) is a highly contagious infectious disease caused by the newly discovered severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Severe COVID-19 infection causes complications in the respiratory tract, which results in pulmonary failure, thus requiring prolonged mechanical ventilation (MV). An increase in the number of patients with COVID-19 poses numerous challenges to the healthcare system, including the shortage of MV facilities. Despite continued efforts to improve COVID-19 diagnosis and treatment, no study has established a reliable predictive model for the risk assessment of deteriorating COVID-19 cases. MATERIALS AND METHODS: We extracted the expression profiles and clinical data of the GSE157103, GSE116560 and GSE21802 cohorts from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified as the intersection of the resulting differential genes as analysed via limma, edgeR and DESeq2 R packages. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed using the R package 'clusterProfiler'. Variables closely related to MV were examined using univariate Cox regression analysis, and significant variables were subjected to least absolute shrinkage and selection operator regression (LASSO) analysis for the construction of a risk model. Kaplan-Meier analysis and receiver operating characteristic (ROC) curves were generated to verify the predictive values of the risk model. RESULTS: We identified 198 unigenes that were differentially expressed in COVID-19 samples. Moreover, a five-gene signature (BTN3A1, GPR35, HAAO, SLC2A6 and TEX2) was constructed to predict the ventilator-free days of patients with COVID-19. In our study, we used the five-gene signature to calculate the risk score (MV score) for each patient. The results revealed a statistical correlation between the MV score and the scores of the Acute Physiology and Chronic Health Evaluation and Sequential Organ Failure Assessment of patients with COVID-19. Kaplan-Meier analysis revealed that the number of ventilator-free days was significantly reduced in the low-MVscore group compared to the high-MVscore group. The ROC curves revealed that our model had a good performance, and the areas under the ROC curve were 0.93 (3-week ROC) and 0.97 (4-week ROC). The 'Limma' package analysis revealed 71 upregulated genes and 59 downregulated genes in the high-MV score group compared to the low-MV score group. These DEGs were mainly enriched in cytokine signalling in immune system and cellular response to cytokine stimulus. CONCLUSIONS: This study identified a five-gene signature that can predict the length of ventilator-free days for patients with COVID-19.

Effects of SARS-CoV-2 Membrane Glycoprotein on Host Pre-mRNA 3 ' Processing

Objective To investigate the effect of SARS-CoV-2 membrane protein on the processing of the 3' untranslated region (UTR) of the mRNA precursor (pre-mRNA) in host cells. Methods Based on the cell model of human lung epithelial cells A549, over-expression of the SARS-CoV-2 membrane protein was performed. The RNA-Seq high-throughput sequencing technique and bioinformatics methods was employed to analyze the systematic characterization of alternative polyadenylation (APA) events in host cells. Genes with significant APA events were uploaded to the Metascape database for functional enrichment analysis. In addition, alternative 3'UTR length of genes with APA events was verified by RT-qPCR. Then the target protein expression level was detected by Western blot. Results A total of 813 genes that were significant dynamic APA events in host cells that overexpressed SARS-CoV-2 membrane protein. These genes were enriched in cell biologicial processes such as the mitotic cell cycle and regulation of cellular response to stress. We further screened AKT1, which encodes a critical regulator involved in the above biological process, showing a 3'UTR lengthening in IGV software. RT-qPCR verified the trend of 3'UTR length changes of AKT1. Western blot showed the increased level of phosphorylated AKT1 protein in over-expressed group of M protein. Conclusion SARS-CoV-2 membrane protein potentially affects the 3' processing of host pre-mRNAs. AKT1, which is involved in a variety of viral biological processes, with 3'UTR lengthening, and its protein function was activated intracellularly.

[Analysis of non-bacterial respiratory pathogens in children in Ningbo City from 2019 to 2021].

Objective: To investigate the distribution characteristics of respiratory non-bacterial pathogens in children in Ningbo from 2019 to 2021. Methods: A retrospective analysis was performed on 23 733 children with respiratory tract infection who visited the department of pediatrics of Ningbo Women and Children's Hospital from July 2019 to December 2021. There were 13 509 males (56.92%) and 10 224 females (43.08%), with an age range of 1 day to 18 years old. There were 981 cases in the neonatal group (younger than 1 month old), 5 880 cases in the infant group (1 month to younger than 1 year old), 6 552 cases in the toddler group (1 to younger than 3 years old), 7 638 cases in the preschool group (3 to younger than 7 years old), and 2 682 cases in the school-age group (7 to 18 years old). Thirteen respiratory pathogens were detected by multiple polymerase chain reaction (PCR) based on capillary electrophoresis, and SPSS 23.0 software was used for statistical analysis of the results, the count data were expressed as percentages, and the χ2 test was used for comparison between groups. Results: Of the 23 733 specimens, 13 330 were positive for respiratory pathogens, with a total positive rate of 56.17%. The positive rates of human rhinovirus (HRV) 24.05% (5 707/23 733), human respiratory syncytial virus (HRSV) 10.45% (2 480/2 3733) and mycoplasma pneumoniae (Mp) 7.03% (1 668/23 733) were in the first three. The positive rates of pathogens in the male and female children were 57.47% (7 763/13 509) and 54.45% (5 567/10 224), respectively, and the difference was statistically significant (χ2=21.488, P<0.001). The positive rates in the neonatal group, infant group, toddler group, preschool group, and school-age group were 31.80% (312/981), 54.71% (3 217/5 880), 63.23% (4 143/6 552), 59.83% (4 570/7 638), 40.57% (1 088/2 682), respectively, and the difference among the groups was statistically significant (χ2=681.225, P<0.001). The single infection rate was 47.43% (11 256/23 733), the mixed infection rate of two or more pathogens was 8.74% (2 074/23 733), most of which were mixed infections of two pathogens. HRV, HADV, HCOV, Ch disseminated in the whole year. HRSV, HMPV, Boca, HPIV occurred mostly in fall and winter. The positive rates of FluA, FluB, Mp were at a low level after the corona virus disease 2019 (COVID-19) epidemic (2020 and 2021). The positive rates of FluA, H1N1, H3N2, FluB, HADV, Mp in 2020 were significantly lower than in 2019 (P<0.05). The positive rates of HPIV, HRV, HCOV, Ch in 2020 were significantly higher than in 2019 (P<0.05). The positive rates of FluA, H1N1, H3N2, HPIV, HCOV, Mp, Ch in 2021 were significantly lower than in 2020 (P<0.05). The positive rates of Boca, HMPV, HRSV in 2021 were significantly higher than in 2020 (P<0.05). Conclusion: From 2019 to 2021, the main non-bacterial respiratory pathogens of children in Ningbo City were Mp and HRV, and the detection rates of respiratory pathogens varied among different ages, seasons and genders.

COVID-VIT: Classification of Covid-19 from 3D CT chest images based on vision transformer model

This paper presents an explainable deep learning network to classify COVID from non-COVID based on 3D CT lung images. It applies a subset of the data for MIA-COV19 challenge through the development of 3D form of Vision Transformer deep learning architecture. The data comprise 1924 subjects with 851 being diagnosed with COVID, among them 1,552 being selected for training and 372 for testing. While most of the data volume are in axial view, there are a number of subjects' data are in coronal or sagittal views with 1 or 2 slices are in axial view. Hence, while 3D data based classification is investigated, in this competition, 2D axial-view images remains the main focus. Two deep learning methods are studied, which are vision transformer (ViT) based on attention models and DenseNet that is built upon conventional convolutional neural network (CNN). Initial evaluation results indicates that ViT performs better than DenseNet with F1 scores being 0.81 and 0.72 respectively. (Codes are available at GitHub at https://github.com/xiaohong1/COVID-ViT). This paper illustrates that vision transformer performs the best in comparison to the other current state of the art approaches in classification of COVID from CT lung images. © 2022 IEEE.

A Non-autonom SIR Model in Epidemiology

With the outbreak of COVID-19, the modelling of epidemic spread has once again become highly important. This paper introduces an epidemic spreading model with a changing infection rate. This model extends the traditional SIR (Susceptible – Infected – Removed) model. The SIR model is a dynamic model which divides individuals into 3 groups: susceptible, infected, and removed (including recovered and died). Individuals in each group have a constant proportion to change to the next group. This paper assumes the infection rate is dependent on the development cycle of the virus, which can vary in the different periods since being infected, instead of constants. This makes the differential equations a non-autonomous model. This paper works on how to fit the function of the infection rate and solve the equations. This paper uses Burr distribution which has 3 unknown parameters as the function of infection rate, and then discusses about two different methods to get these parameters—the least-squares method and the maximum likelihood estimation. As a numerical experiment of this model, this paper uses the data of COVID-19 in Ireland to make predictions and compare with the traditional SIR model. The non-autonomous model in this paper shows better performance than the traditionary SIR model. This new model might be potential in further epidemic simulation, and it is not hard to be combined with other extensions of the SIR model. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Network Public Opinion Detection During the Coronavirus Pandemic: A Short-Text Relational Topic Model

Online social media provides rich and varied information reflecting the significant concerns of the public during the coronavirus pandemic. Analyzing what the public is concernedwith from social media information can support policy-makers to maintain the stability of the social economy and life of the society. In this article, we focus on the detection of the network public opinions during the coronavirus pandemic. We propose a novel Relational Topic Model for Short texts (RTMS) to draw opinion topics from social media data. RTMS exploits the feature of texts in online social media and the opinion propagation patterns among individuals. Moreover, a dynamic version of RTMS (DRTMS) is proposed to capture the evolution of public opinions. Our experiment is conducted on a real-world dataset which includes 67,592 comments from 14,992 users. The results demonstrate that, compared with the benchmark methods, the proposed RTMS and DRTMS models can detect meaningful public opinions by leveraging the feature of social media data. It can also effectively capture the evolution of public concerns during different phases of the coronavirus pandemic.

A Survey of Machine Learning Technologies for COVID-19 Pandemic

Intelligent technologies including machine learning and artificial intelligence are playing significant roles in human's battle against COVID-19 pandemic. Machine learning enables the machine to learn and improve on its own without being programmed in detail. Machine learning has now penetrated into many fields to help fight the epidemic. However, a specific and representative review of the contributions of machine learning is currently lacking. The purpose of this paper is to summarize several machine learning applications against COVID-19 including: i) predicting confirmed cases and trend, ii) classifying and diagnosing using ML-based images, and iii) managing medical resources. A database related to machine learning Technologies for COVID-19 is created. Moreover, a concise review is finished on the collected information by evaluating the different uses of machine learning against COVID-19. We also assemble researches in the present COVID-19 literature focused on ML-based methods in order to demonstrate a profound insight into COVID-19 related topics. Our discoveries emphasize crucial variables and available COVID-19 resources that facilitate clinical and translational research. © 2022 IEEE.

[Psychological and behavioral functioning of children and adolescents during long-term home-schooling].

Objective: To study the characteristics and risk factors of psychological and behavioral problems of children and adolescents of different ages and genders in long-term home-schooling during the coronavirus disease-2019 pandemic. Further, to provide scientific basis for more targeted psychological intervention and coping strategies in the future. Methods: A cross-sectional survey using an online questionnaire was conducted on students aged 6-16 years old in five representative cities of North (Beijing), East (Shanghai), West (Chongqing), South (Guangzhou) and Middle (Wuhan) in China. In this study, the social behavior and psychological abnormalities which was defined as the positive of any dimension were investigated in multiple dimensions during long-term home-schooling. The influencing factors of psycho-behavioral problems were analyzed by Logistic regression, and the confounding factors were corrected with graded multivariable adjustment. Results: A total of 6 906 valid questionnaires were collected including 3 592 boys and 3 314 girls, of whom 3 626 were children (6-11 years old) and 3 280 were adolescents (12-16 years old). The positive detection rate of psychosocial-behavioral problems were 13.0% (900/6 906) totally, 9.6% (344/3 592) in boys and 16.8% (556/3 314) in girls respectively, and 7.3%(142/1 946) in boys aged 6-11, 14.0%(235/1 680) in girls aged 6-11, 12.3%(202/1 646) in boys aged 12-16, 19.6%(321/1 634) in girls aged 12-16 respectively. There were significant differences between the psychological problems group and the non-psychological problems group in gender, parent-offspring conflict, number of close friends, family income change, sedentary time, homework time, screen exposure time, physical activity, dietary problems (χ²=78.851, 285.264, 52.839, 26.284, 22.778, 11.024, 10.688, 36.814, 70.982, all P<0.01). The most common symptoms in boys aged 6-11 years were compulsive activity, schizoid and depression, in girls aged 6-11 years were schizoid/compulsive activity, hyperactivity and social withdrawal, in boys aged 12-16 years were hyperactivity, compulsive activity and aggressive behavior, and in girls aged 12-16 years were schizoid, anxiety/compulsive activity and depression/withdrawal, respectively. After graded multivariable adjustment, besides the common risk factors, homework time and online study time were the risk factors of 6-11 years old groups [boys OR(95%CI): 1.750 (1.32-2.32), 1.214(1.00-1.47), girls: 1.579(1.25-1.99), 1.222(1.05-1.42), all P<0.05], videogames time were the risk factors of 12-16 years old groups [ boys: 2.237 (1.60-3.13), girls: 1.272 (1.00-1.61), all P<0.05]. Conclusions: Some children and adolescents may have psychological and behavioral problems during long-term home-schooling. The psychological and behavioral manifestations differed in age and gender subgroups, which deserve special attention in each subgroups. Schools, families and specialists should actively provide precise psychological support and comprehensive intervention strategies according to special features and risk factors.

Investors Embrace Gender Diversity, Not Female CEOs: The Role of Gender in Startup Fundraising

The allocation of venture capital is one of the primary factors determining who takes products to market, which startups succeed or fail, and as such who gets to participate in the shaping of our collective economy. While gender diversity contributes to startup success, most funding is allocated to male-only entrepreneurial teams. In the wake of COVID-19, 2020 is seeing a notable decline in funding to female and mixed-gender teams, giving raise to an urgent need to study and correct the longstanding gender bias in startup funding allocation. We conduct an in-depth data analysis of over 48,000 companies on Crunchbase, comparing funding allocation based on the gender composition of founding teams. Detailed findings across diverse industries and geographies are presented. Further, we construct machine learning models to predict whether startups will reach an equity round, revealing the surprising finding that the CEO’s gender is the primary determining factor for attaining funding. Policy implications for this pressing issue are discussed. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

Outcomes for hospitalized cancer patients with COVID-19 during the height of pandemic in New York City

Background: Several reports have suggested that cancer patients are at increased risk for contracting severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and suffering worse coronavirus disease 2019 (COVID-19) outcomes. However, little is known about the impact of cancer status on presentation and outcome. Here, we report on the association between cancer status and survival in hospitalized patients who tested positive for SARSCoV- 2 during the height of pandemic in New York City. Methods: Of the 6,724 patients who were hospitalized at NYU Langone Health (3/16/20 -7/31/20) and tested positive for SARS-CoV-2, 580 had either active cancer (n = 221) or a history of cancer (n = 359). Patients were classified as having active malignancy if they either received treatment within six months of their COVID-19 diagnosis or they had measurable disease documented at the time of their hospitalization. Patients were categorized as having a history of cancer if there was no evidence of measurable disease or there were no treatments administered within six months of their COVID-19 diagnosis. We compared the baseline clinicodemographic characteristics and hospital courses of the two groups, and the relationship between cancer status and the rate of admission to the intensive care unit (ICU), use of invasive mechanical ventilation (IMV), and all-cause mortality. Results: There was no differences between the two groups in their baseline laboratory results associated with COVID- 19 infection, incidence of venous thromboembolism, or incidence of severe COVID- 19. Active cancer status was not associated with the rate of ICU admission (P =0.307) or use of IMV (P = 0.236), but was significantly associated with worse all-cause mortality in both univariate and multivariate analysis with ORs of 1.48 (95% CI: 1.04-2.09;P = 0.028) and 1.71 (95% CI: 1.12- 2.63;P = 0.014), respectively. Conclusions: Active cancer patients had worse survival outcomes compared to patients with a history of cancer despite similar COVID-19 disease characteristics in the two groups. Our data suggest that cancer care should continue with minimal interruptions during the pandemic to bring about response and remission as soon as possible. Additionally, these findings support the growing body of evidence that malignancy portends worse COVID-19 prognosis, and demonstrate that the risk may even apply to those without active disease.

First experience with hyperpolarized 129xe imaging in a recovered covid-19 patient

Introduction: There are a growing number of reports of persistently reduced exercise capacities, dyspnea or cough in a small fraction of Covid-19 survivors, suggesting ongoing impaired lung function long after the acute infection has resolved. The cause of these symptoms is unclear, though they likely originate in subtle damage to alveolar septa or vasculature. Here, we present the case of a patient with persistent post-COVID-19 symptoms who was evaluated with hyperpolarized xenon-129 MRI methods, which are sensitive to both ventilation and exchange in both non-specific tissue-plasma and red-blood-cell bound compartments in the lungs. Case: A 58-year-old never-smoker female patient was diagnosed COVID-19 positive in August 2020. She continued to experience nonspecific symptoms of fatigue, pins-and-needles in the feet, dyspnea, and daily productive cough (green, non-bloody sputum). Chest x-ray showed clear lungs without focal consolidation, pleural effusion, or pneumothorax. The subject underwent xenon-129 MR imaging on December 11, 2020 using a multi-breath scheme, in which sets of 6 ad libitum breaths containing 50mL of hyperpolarized xenon-129 (balance room air) were followed by four breaths of room air, and that 10-breath sequence was repeated until the polarized xenon-129 gas supply was exhausted. As shown in Figure 1, ventilated lung volumes are visually patchy, with heterogeneity corresponding to lobar structures or segmental and subsegmental volumes that are likely fed by airways with varying degrees of blockage. This is consistent with the persistent sputum production experienced by the patient. Further, saturation pulses at the frequency of hemoglobin-bound and tissue-plasma xenon-129 resonances selectively destroy signal in their respective compartments, which is subsequently exchanged to the gas phase. Compared to a healthy volunteer, the fractional depolarization achieved when applying identical saturation pulses is reduced by an average of approximately 40% in the patient. The response to saturation pulses also exhibits significant spatial heterogeneity. Discussion: Although a single case study is unable to determine the origin of alterations seen in a recovered COVID-19 patient, these changes are consistent with an overall reduction in the rate of gas exchange into dissolved compartments, as well as with a somewhat heterogeneous pattern of ventilation characteristic of mild obstructive disease. Further studies will be required to determine if these changes are associated with severe or persistent morbidity, and if correspondence to an age-matched healthy cohort increases as recovery continues.

Community Mitigation: A Data-driven System for COVID-19 Risk Assessment in a Hierarchical Manner

The fast evolving and deadly outbreak of coronavirus disease (COVID-19) has posed grand challenges to human society. To slow the spread of virus infections and better respond with actionable strategies for community mitigation, leveraging the large-scale and real-time pandemic related data generated from heterogeneous sources (e.g., disease related data, demographic data, mobility data, and social media data), in this work, we propose and develop a data-driven system (named α-satellite), as an initial offering, to provide real-time COVID-19 risk assessment in a hierarchical manner in the United States. More specifically, given a location (either user input or automatic positioning), the system will automatically provide risk indices associated with the specific location, the county that location is in and the state as a whole to enable people to select appropriate actions for protection while minimizing disruptions to daily life to the extent possible. In α-satellite, we first construct an attributed heterogeneous information network (AHIN) to model the collected multi-source data in a comprehensive way;and then we utilize meta-path based schemes to model both vertical and horizontal information associated with a given location (i.e., point of interest, POI);finally we devise a novel heterogeneous graph neural network to aggregate its neighborhood information to estimate the risk of the given POI in a hierarchical manner. To comprehensively evaluate the performance of α-satellite in real-time COVID-19 risk assessment, a set of studies are first performed to validate its utility;based on a real-world dataset consisting of 6,538 annotated POIs, the experimental results show that α-satellite achieves the area of under curve (AUC) of 0.9378, which outperforms the state-of-the-art baselines. After we launched the system for public tests, it had attracted 51,190 users as of May 30. Based on the analysis of its large-scale users, we have a key finding that people from more severe regions (i.e., with larger numbers of COVID-19 cases) have stronger interests using the system for actionable information. Our system and generated benchmark datasets have been made publicly accessible through our website. © 2020 ACM.