Phone No: +86 18071093208 Word count of abstract: 215 Word count of text: 2483 Highlights  The immune status is significantly different between severe and non-severe COVID-19  The decrease of T lymphocyte correlated with the course of patients with COVID-19  The level of T lymphocyte is an indicator for severity and prognosis of COVID-19Abstract: Objectives: To explore the clinical course and its dynamic features of immune status in COVID-19 patients and find predictors correlated with severity and prognosis of COVID-19. Methods: The electronic medical records of 204 patients with COVID-19 pneumonia confirmed by nucleic acid testing were retrospectively collected and analyzed. Results: All Patients were divided into severe (69) and non-severe group (135). Lymphocyte subsets count, including CD3+ T cell, CD4+ T cell, CD8+ T cell, B cell (CD19+) and NK J o u r n a l P r e -p r o o f cell (CD16+56+), were significantly lower in severe group (P＜0.001). The dynamic levels of T lymphocytes in severe group were significantly lower from disease onset, but in the improved subgroup the value of T lymphocyte began to increase after about 15-day treatment and finally returned to the normal level. The cut-off value of the counts of CD3+ (576), CD4+ (391) and CD8+ (214) T cell were calculated and indicated significantly high sensitivity and specificity for severity of COVID-19. Conclusion: Our results shown that the decrease of CD3+, CD4+ and CD8+ T lymphocyte correlated with the course of patients with COVID-19 pneumonia, especially in severe cases. The level of T lymphocyte could be used as an indicator for prediction of severity and prognosis of patients with COVID-19 pneumonia. The application of glucocorticoid should be cautious in severe cases. IgE, IU/mL <100 23.5 (17.3-80.8) 21.9 (17.3-84.3) 30.85 (17.3-68.9) 0.707J o u r n a l P r e -p r o o f

Aimed to characterize the CT imaging and clinical course of asymptomatic cases with COVID-19 pneumonia. Methods: Asymptomatic cases with COVID-19 pneumonia confirmed by SARS-COV-2 nucleic acid testing in Renmin Hospital of Wuhan University were retrospectively enrolled. The characteristics of CT imaging and clinical feature were collected and analyzed. Results: 58 asymptomatic cases with COVID-19 pneumonia admitted to our hospital between Jan 1, 2020 and Feb 23, 2020 were enrolled. All patients had history of exposure to SARS-CoV-2. On admission, patients had no symptoms and laboratory findings were normal. The predominant feature of CT findings in this cohort was ground glass opacity (GGO) (55, 94.8%) with peripheral (44, 75.9%) distribution, unilateral location (34, 58.6%) and mostly involving one or two lobes (38, 65.5%), often accompanied by characteristic signs. After short-term follow-up, 16 patients (27.6%) presented symptoms with lower lymphocyte count and higher CRP, mainly including fever, cough and fatigue. The evolution of lesions on CT imaging were observed in 10 patients (17.2%). The average days of hospitalization was19.80 ±10.82 days, and was significantly longer in progression patients (28.60 ±7.55 day). Conclusion: CT imaging of asymptomatic cases with COVID-19 pneumonia has definite characteristics. Since asymptomatic infections as "covert transmitter", and some patients can progress rapidly in the short term. It is essential to pay attention to the surveillance of asymptomatic patients with COVID-19. CT scan has great value in screening and detecting patients with COVID-19 pneumonia, especially in the highly suspicious, asymptomatic cases with negative nucleic acid testing.

Mutations in the SCN1A gene have been identified in epilepsy patients with widely variable phenotypes and modes of inheritance and in asymptomatic carriers. This raises challenges in evaluating the pathogenicity of SCN1A mutations. We systematically reviewed all SCN1A mutations and established a database containing information on functional alterations. In total, 1,257 mutations have been identified, of which 81.8% were not recurrent. There was a negative correlation between phenotype severity and missense mutation frequency. Further analyses suggested close relationships among genotype, functional alteration, and phenotype. Missense mutations located in different sodium channel regions were associated with distinct functional changes. Missense mutations in the pore region were characterized by the complete loss of function, similar to haploinsufficiency. Mutations with severe phenotypes were more frequently located in the pore region, suggesting that functional alterations are critical in evaluating pathogenicity and can be applied to patient management. A negative correlation was found between phenotype severity and familial incidence, and incomplete penetrance was associated with missense and splice site mutations, but not truncations or genomic rearrangements, suggesting clinical genetic counseling applications. Mosaic mutations with a load of 12.5-25.0% were potentially pathogenic with low penetrance, suggesting the need for future studies on less pathogenic genomic variations.

We present a novel algorithm for point cloud segmentation. Our approach transforms unstructured point clouds into regular voxel grids, and further uses a kernel-based interpolated variational autoencoder (VAE) architecture to encode the local geometry within each voxel. Traditionally, the voxel representation only comprises Boolean occupancy information which fails to capture the sparsely distributed points within voxels in a compact manner. In order to handle sparse distributions of points, we further employ radial basis functions (RBF) to compute a local, continuous representation within each voxel. Our approach results in a good volumetric representation that effectively tackles noisy point cloud datasets and is more robust for learning. Moreover, we further introduce group equivariant CNN to 3D, by defining the convolution operator on a symmetry group acting on Z 3 and its isomorphic sets. This improves the expressive capacity without increasing parameters, leading to more robust segmentation results. We highlight the performance on standard benchmarks and show that our approach outperforms state-of-the-art segmentation algorithms on the ShapeNet and S3DIS datasets.