The novel coronavirus (2019-nCoV) is spreading very fast in Hubei Province of China. As of February 14, 2020, 51,986 confirmed cases (including laboratory-confirmed cases and clinically-confirmed cases) were reported in Hubei Province, and 1,318 of them died. Respiratory droplets and contact transmission are considered to be the most important routes of transmission of 2019-nCoV, but do not fully account for the occurrence of all coronavirus disease 2019 (COVID-19) cases, previously known as novel coronavirus pneumonia (NCP), and the reasons for the rapid spread of this virus (1).In Biosafety Level 3 (BSL-3) Laboratory of the National Institute for Viral Disease Control and Prevention, Vero cells were used for viral isolation from stool samples of COVID-19 patients sent by Heilongjiang CDC. A 2019-nCoV strain was isolated from a stool specimen of a laboratory-confirmed COVID-19 severe pneumonia case, who experienced onset on January 16, 2020 and was sampled on February 1, 2020. The interval between sampling and onset was 15 days. The full-length genome sequence indicated that the virus had high-nucleotide similarity (99.98%) to that of the first isolated novel coronavirus isolated from Wuhan, China (Figure 1). In the Vero cells, viral particles with typical morphology of a

Background and Aims The evolution and clinical significance of abnormal liver chemistries and the impact of hepatitis B infection on outcome in patients with COVID-19 is not well characterized. This study aimed to explore these issues. Methods This large retrospective cohort study included 2073 patients with COVID-19 having definite outcomes in Wuhan, China. Longitudinal liver function tests were conducted and determined their associated factors and death risk by multivariate regression analyses. A prognostic nomogram was formulated to predict the survival of patients with COVID-19. The characteristics of liver abnormalities and outcomes of patients with COVID-19 with and without hepatitis B were compared after 1:3 propensity score matching. Results Of the 2073 patients, 1282 (61.8%) had abnormal liver chemistries during hospitalization, and 297 (14.3%) had a liver injury. The mean levels of AST and D-Bil increased early after symptom onset in deceased patients and showed disparity compared with that in discharged patients throughout the clinical course of the disease. Abnormal admission AST (adjusted hazard ratio [HR]: 1.39, 95%CI: 1.04-1.86, P =0.027) and D-Bil (adjusted HR: 1.66, 95%CI: 1.22-2.26, P =0.001) levels were independent risk factors for mortality due to COVID-19. A nomogram was established based on the results of multivariate analysis and showed sufficient discriminatory power and good consistency between the prediction and the observation. HBV infection in patients did not increase the risk of COVID-19-associated poor outcomes. Conclusions Abnormal AST and D-Bil levels at admission were independent predictors of COVID-19 mortality. Therefore, monitoring liver chemistries, especially AST and D-Bil levels, in hospitalized patients with COVID-19, is necessary.

NCT01770431; Post-results.

[1] Recent theoretical and empirical studies show that the generalization ability of artificial neural networks can be improved by combining several artificial neural networks in redundant ensembles. In this paper, a review is given of popular ensemble methods. Six approaches for creating artificial neural network ensembles are applied in pooled flood frequency analysis for estimating the index flood and the 10-year flood quantile. The results show that artificial neural network ensembles generate improved flood estimates and are less sensitive to the choice of initial parameters when compared with a single artificial neural network. Factors that may affect the generalization of an artificial neural network ensemble are analyzed. In terms of the methods for creating ensemble members, the model diversity introduced by varying the initial conditions of the base artificial neural networks to reduce the prediction error is comparable with more sophisticated methods, such as bagging and boosting. When the same method for creating ensemble members is used, combining member networks using stacking is generally better than using simple averaging. An ensemble size of at least 10 artificial neural networks is suggested to achieve sufficient generalization ability. In comparison with parametric regression methods, properly designed artificial neural network ensembles can significantly reduce the prediction error.