This thesis explores to what extent Multinomial Naive Bayes (MNB) and Support Vector Machines (SVM) classifiers can be used to determine the polarity of news, specifically the news coverage of Sweden by the Russian state-funded news outlets RT and Sputnik. Three experiments are conducted.  In the first experiment, an MNB and an SVM classifier are trained with the Large Movie Review Dataset (Maas et al., 2011) with a varying number of samples to determine how training data size affects classifier performance.  In the second experiment, the classifiers are trained with 300 positive, negative, and neutral news articles (Agarwal et al., 2019) and tested on 95 RT and Sputnik news articles about Sweden (Bengtsson, 2019) to determine if the domain specificity of the training data outweighs its limited size.  In the third experiment, the movie-trained classifiers are put up against the domain-specific classifiers to determine if well-trained classifiers from another domain perform better than relatively untrained, domain-specific classifiers.  Four different types of feature sets (unigrams, unigrams without stop words removal, bigrams, trigrams) were used in the experiments. Some of the model parameters (TF-IDF vs. feature count and SVM’s C parameter) were optimized with 10-fold cross-validation.  Other than the superior performance of SVM, the results highlight the need for comprehensive and domain-specific training data when conducting machine learning tasks, as well as the benefits of feature engineering, and to a limited extent, the removal of stop words. Interestingly, the classifiers performed the best on the negative news articles, which made up most of the test set (and possibly of Russian news coverage of Sweden in general).

Indiana University-Purdue University Indianapolis (IUPUI) / Mammalian target of rapamycin (mTOR) is a key regulator of protein synthesis. Dysregulation of mTOR signaling occurs in many human cancers and its inhibition causes arrest at the G1 cell cycle stage. However, mTOR’s impact on mitosis (M-phase) is less clear. Here, suppressing mTOR activity impacted the G2-M transition and reduced levels of M-phase kinase, Aurora A. mTOR inhibitors did not affect Aurora A mRNA levels. However, translational reporter constructs showed that mRNA containing a short, simple 5’-untranslated region (UTR), rather than a complex structure, is more responsive to mTOR inhibition. mTOR inhibitors decreased Aurora A protein amount whereas blocking proteasomal degradation rescues this phenomenon, revealing that mTOR affects Aurora A protein stability. Inhibition of protein phosphatase, PP2A, a known mTOR substrate and Aurora A partner, restored mTOR-mediated Aurora A abundance. Finally, a non-phosphorylatable Aurora A mutant was more sensitive to destruction in the presence of mTOR inhibitor. These data strongly support the notion that mTOR controls Aurora A destruction by inactivating PP2A and elevating the phosphorylation level of Ser51 in the “activation-box” of Aurora A, which dictates its sensitivity to proteasomal degradation. In summary, this study is the first to demonstrate that mTOR signaling regulates Aurora-A protein expression and stability and provides a better understanding of how mTOR regulates mitotic kinase expression and coordinates cell cycle progression. The involvement of mTOR signaling in the regulation of cell migration by its upstream activator, Rheb, was also examined. Knockdown of Rheb was found to promote F-actin reorganization and was associated with Rac1 activation and increased migration of glioma cells. Suppression of Rheb promoted platelet-derived growth factor receptor (PDGFR) expression. Pharmacological inhibition of PDGFR blocked these events. Therefore, Rheb appears to suppress tumor cell migration by inhibiting expression of growth factor receptors that in turn drive Rac1-mediate actin polymerization.

The goal of this thesis is to improve our knowledge on energetic particle precipitation into the Earth’s atmosphere from the thermosphere to the surface. The particles origin from the Sun or from temporarily trapped populations inside the magnetosphere. The best documented influence of solar (high-) energetic particles on the atmosphere is the Ozone depletion in high latitudes, attributed to the generation of HOx and NOx by precipitating particles (Crutzen et al., 1975; Solomon et al., 1981; Reid et al., 1991). In addition Callis et al. (1996b, 2001) and Randall et al. (2005, 2006) point out the importance of low-energetic precipitating particles of magnetospheric origin, creating NOx in the lower thermosphere, which may be transported downwards where it also contributes to Ozone depletion. The incoming particle flux is dramatically changing as a function of auroral/geomagnetical activity and in particular during solar particle events. As a result, the degree of ionization and the chemical composition of the atmosphere are substantially affected by the state of the Sun. Therefore the direct energetic or dynamical influences of ions on the upper atmosphere depend on solar variability at different time scales. Influences on chemistry have been considered so far with simplified precipitation patterns, limited energy range and restrictions to certain particle species, see e.g. Jackman et al. (2000); Sinnhuber et al. (2003b, for solar energetic protons and no spatial differentiation), and Callis et al. (1996b, 2001, for magnetospheric electrons only). A comprehensive atmospheric ionization model with spatially resolved particle precipitation including a wide energy range and all main particle species as well as a dynamic magnetosphere was missing. In the scope of this work, a 3-D precipitation model of solar and magnetospheric particles has been developed. Temporal as well as spatial ionization patterns will be discussed. Apart from that, the ionization data are used in different climate models, allowing (a) simulations of NOx and HOx formation and transport, (b) comparisons to incoherent scatter radar measurements and (c) inter-comparison of the chemistry part in different models and comparison of model results to MIPAS observations. In a bigger scope the ionization data may be used to better constrain the natural sources of climate change or consequences for atmospheric dynamics due to local temperature changes by precipitating particles and their implications for chemistry. Thus the influence of precipitating energetic particles on the composition and dynamics of the atmosphere is a challenging issue in climate modeling. The ionization data is available online and can be adopted automatically to any user specific model grid.