COVID-19 infection inference with graph neural networks.

Infectious diseases spread rapidly, and epidemiological surveys are vital to detect high-risk transmitters and reduce transmission rates. To enhance efficiency and reduce the burden on epidemiologists, an automatic tool to assist with epidemiological surveys is necessary. This study aims to develop an automatic epidemiological survey to predict the influence of COVID-19-infected patients on future additional infections. To achieve this, the study utilized a dataset containing interaction information between confirmed cases, including contact order, contact times, and movement routes, as well as individual properties such as symptoms. Graph neural networks (GNNs) were used to incorporate interaction information and individual properties. Two variants of GNNs, graph convolutional and graph attention networks, were utilized, and the results showed that the graph-based models outperformed traditional machine learning models. For the area under the curve, the 2nd, 3rd, and 4th order spreading predictions showed higher performance by 0.200, 0.269, and 0.190, respectively. The results show that the contact information of an infected person is crucial data that can help predict whether that person will affect future infections. Our findings suggest that incorporating the relationships between an infected person and others can improve the effectiveness of an automatic epidemiological survey.

Comparative Effectiveness of Intracerebroventricular, Intrathecal, and Intranasal Routes of AAV9 Vector Administration for Genetic Therapy of Neurologic Disease in Murine Mucopolysaccharidosis Type I.

Mucopolysaccharidosis type I (MPS I) is an inherited metabolic disorder caused by deficiency of the lysosomal enzyme alpha-L-iduronidase (IDUA). The two current treatments [hematopoietic stem cell transplantation (HSCT) and enzyme replacement therapy (ERT)], are insufficiently effective in addressing neurologic disease, in part due to the inability of lysosomal enzyme to cross the blood brain barrier. With a goal to more effectively treat neurologic disease, we have investigated the effectiveness of AAV-mediated IDUA gene delivery to the brain using several different routes of administration. Animals were treated by either direct intracerebroventricular (ICV) injection, by intrathecal (IT) infusion into the cerebrospinal fluid, or by intranasal (IN) instillation of AAV9-IDUA vector. AAV9-IDUA was administered to IDUA-deficient mice that were either immunosuppressed with cyclophosphamide (CP), or immunotolerized at birth by weekly injections of human iduronidase. In animals treated by ICV or IT administration, levels of IDUA enzyme ranged from 3- to 1000-fold that of wild type levels in all parts of the microdissected brain. In animals administered vector intranasally, enzyme levels were 100-fold that of wild type in the olfactory bulb, but enzyme expression was close to wild type levels in other parts of the brain. Glycosaminoglycan levels were reduced to normal in ICV and IT treated mice, and in IN treated mice they were normalized in the olfactory bulb, or reduced in other parts of the brain. Immunohistochemical analysis showed extensive IDUA expression in all parts of the brain of ICV treated mice, while IT treated animals showed transduction that was primarily restricted to the hind brain with some sporadic labeling seen in the mid- and fore brain. At 6 months of age, animals were tested for spatial navigation, memory, and neurocognitive function in the Barnes maze; all treated animals were indistinguishable from normal heterozygous control animals, while untreated IDUA deficient animals exhibited significant learning and spatial navigation deficits. We conclude that IT and IN routes are acceptable and alternate routes of administration, respectively, of AAV vector delivery to the brain with effective IDUA expression, while all three routes of administration prevent the emergence of neurocognitive deficiency in a mouse MPS I model.

Homogeneous dispersion of TiC nano particles in a cast carbon steel matrix.

Metal matrix nano-composites (MMNCs) (metal matrix with nano-sized ceramic particles) can be of great significance because of their high performance and thus it would be advantageous to produce as-cast bulk MMNCs. However, it is so difficult to disperse nano-sized ceramic particles uniformly in molten metal. In this study, carbon steel matrix composites with a homogeneous dispersion of TiC nano particles were fabricated by conventional liquid metal casting method. In order to get highly wettable nano-sized TiC ceramic particles, the micro-sized (approximately 10 m) TiC particles were first mechanically milled (MMed) by Cu in a high-energy ball mill machine (MMed TiC/Cu), and then mixed with Sn powders to obtain better wettability, as this lowered the surface tension of the carbon steel melt. According to OM images, an addition of MMed TiC/Cu-Sn mixed powders favorably disperses the TiC nano particles in the carbon steel matrix. SEM and EDS images revealed that spherical particles with several hundreds of nanometers were distributed uniformly in the carbon steel matrix. It was also found that the grain size refinement of the cast matrix is achieved remarkably when TiC nano particles were added due to the fact that TiC nano particles act as nucleation sites during the solidification process.