Mining biomedical information from scientific literature

The rapid evolution and proliferation of a world-wide computerized network, the Internet, resulted in an overwhelming and constantly growing amount of publicly available data and information, a fact that was also verified in biomedicine. However, the lack of structure of textual data inhibits its direct processing by computational solutions. Information extraction is the task of text mining that intends to automatically collect information from unstructured text data sources. The goal of the work described in this thesis was to build innovative solutions for biomedical information extraction from scientific literature, through the development of simple software artifacts for developers and biocurators, delivering more accurate, usable and faster results. We started by tackling named entity recognition - a crucial initial task - with the development of Gimli, a machine-learning-based solution that follows an incremental approach to optimize extracted linguistic characteristics for each concept type. Afterwards, Totum was built to harmonize concept names provided by heterogeneous systems, delivering a robust solution with improved performance results. Such approach takes advantage of heterogenous corpora to deliver cross-corpus harmonization that is not constrained to specific characteristics. Since previous solutions do not provide links to knowledge bases, Neji was built to streamline the development of complex and custom solutions for biomedical concept name recognition and normalization. This was achieved through a modular and flexible framework focused on speed and performance, integrating a large amount of processing modules optimized for the biomedical domain. To offer on-demand heterogenous biomedical concept identification, we developed BeCAS, a web application, service and widget. We also tackled relation mining by developing TrigNER, a machine-learning-based solution for biomedical event trigger recognition, which applies an automatic algorithm to obtain the best linguistic features and model parameters for each event type. Finally, in order to assist biocurators, Egas was developed to support rapid, interactive and real-time collaborative curation of biomedical documents, through manual and automatic in-line annotation of concepts and relations. Overall, the research work presented in this thesis contributed to a more accurate update of current biomedical knowledge bases, towards improved hypothesis generation and knowledge discovery.

Coupling of magnetic, strain and electric polarization fields in the structure of multiferroic material

The comprehensive study on the coupling of magnetism, electrical polarization and the crystalline lattice with the off-stoichiometric effects in self-doped multiferroic hexagonal h-LuMnxO3±δ (0.92≤x≤1.12) ceramic oxides was carried out for the PhD work. There is a complex coupling of the three ferroic degrees. The cancelation of the magnetic moments of ions in the antiferromagnetic order, electric polarization with specific vortex/antivortex topology and lattice properties have pushed researchers to find out ways to disclose the underlying physics and chemistry of magneto-electric and magneto-elastic couplings of h-RMnO3 multiferroic materials. In this research work, self-doping of Lu-sites or Mn-sites of h-LuMnxO3±δ ceramics prepared via solid state route was done to pave a way for deeper understanding of the antiferromagnetic transition, the weak ferromagnetism often reported in the same crystalline lattices and the ferroelectric properties coupled to the imposed lattice changes. Accordingly to the aim of the PhD thesis, the objectives set for the sintering study in the first chapter on experimental results were two. First, study of sintering off-stoichiometric samples within conditions reported in the bibliography and also extracted from the phase diagrams of the LuMnxO3±δ, with a multiple firings ending with a last high temperature step at 1300ºC for 24 hours. Second, explore longer annealing times of up to 240 hours at the fixed temperature of 1300 ºC in a search for improving the properties of the solid solution under study. All series of LuMnxO3±δ ceramics for each annealing time were characterized to tentatively build a framework enabling comparison of measured properties with results of others available in literature. XRD and Rietveld refinement of data give the evolution the lattice parameters as a function to x. Shrinkage of the lattice parameters with increasing x values was observed, the stability limit of the solid solution being determined by analysis of lattice parameters. The evolution of grain size and presence of secondary phases have been investigated by means of TEM, SEM, EDS and EBSD techniques. The dependencies of grain growth and regression of secondary phases on composition x and time were further characterized. Magnetic susceptibility of samples and magnetic irreversibility were extensively examined in the present work. The dependency of magnetic susceptibility, Neel ordering transition and important magnetic parameters are determined and compared to observation in other multiferroics in the following chapter of the thesis. As a tool of high sensitivity to detect minor traces of the secondary phase hausmannite, magnetic measurements are suggested for cross-checking of phase diagrams. Difficulty of previous studies on interpreting the magnetic anomaly below 43 K in h-RMnO3 oxides was discussed and assigned to the Mn3O4 phase, with supported of the electron microscopy. Magneto-electric coupling where AFM ordering is coupled to dielectric polarization is investigated as a function of x and of sintering condition via frequency and temperature dependent complex dielectric constant measurements in the final chapter of the thesis. Within the limits of solid solubility, the crystalline lattice of off-stoichiometric ceramics was shown to preserve the magneto-electric coupling at TN. It represents the first research work on magneto-electric coupling modified by vacancy doping to author’s knowledge. Studied lattices would reveal distortions at the atomic scale imposed by local changes of x dependent on sintering conditions which were widely inspected by using TEM/STEM methods, complemented with EDS and EELS spectroscopy all together to provide comprehensive information on cross coupling of distortions, inhomogeneity and electronic structure assembled and discussed in a specific chapter. Internal interfaces inside crystalline grains were examined. Qualitative explanations of the measured magnetic and ferroelectric properties were established in relation to observed nanoscale features of h-LuMnxO3±δ ceramics. Ferroelectric domains and topological defects are displayed both in TEM and AFM/PFM images, the later technique being used to look at size, distribution and switching of ferroelectric domains influenced by vacancy doping at the micron scale bridging to complementary TEM studies on the atomic structure of ferroelectric domains. In support to experimental study, DFT simulations using Wien2K code have been carried out in order to interpret the results of EELS spectra of O K-edge and to obtain information on the cation hybridization to oxygen ions. The L3,2 edges of Mn is used to access the oxidation state of the Mn ions inside crystalline grains. In addition, rehybridization driven ferroelectricity is also evaluated by comparing the partial density of states of the orbitals of all ions of the samples, also the polarization was calculated and correlated to the off-stoichiometric effect.