Optimization and Applications of Large-Scale Biomedical Event Networks

The overwhelming amount and unprecedented speed of publication in the biomedical domain make it difficult for life science researchers to acquire and maintain a broad view of the field and gather all information that would be relevant for their research. As a response to this problem, the BioNLP (Biomedical Natural Language Processing) community of researches has emerged and strives to assist life science researchers by developing modern natural language processing (NLP), information extraction (IE) and information retrieval (IR) methods that can be applied at large-scale, to scan the whole publicly available biomedical literature and extract and aggregate the information found within, while automatically normalizing the variability of natural language statements. Among different tasks, biomedical event extraction has received much attention within BioNLP community recently. Biomedical event extraction constitutes the identification of biological processes and interactions described in biomedical literature, and their representation as a set of recursive event structures. The 2009–2013 series of BioNLP Shared Tasks on Event Extraction have given raise to a number of event extraction systems, several of which have been applied at a large scale (the full set of PubMed abstracts and PubMed Central Open Access full text articles), leading to creation of massive biomedical event databases, each of which containing millions of events. Sinece top-ranking event extraction systems are based on machine-learning approach and are trained on the narrow-domain, carefully selected Shared Task training data, their performance drops when being faced with the topically highly varied PubMed and PubMed Central documents. Specifically, false-positive predictions by these systems lead to generation of incorrect biomolecular events which are spotted by the end-users. This thesis proposes a novel post-processing approach, utilizing a combination of supervised and unsupervised learning techniques, that can automatically identify and filter out a considerable proportion of incorrect events from large-scale event databases, thus increasing the general credibility of those databases. The second part of this thesis is dedicated to a system we developed for hypothesis generation from large-scale event databases, which is able to discover novel biomolecular interactions among genes/gene-products. We cast the hypothesis generation problem as a supervised network topology prediction, i.e predicting new edges in the network, as well as types and directions for these edges, utilizing a set of features that can be extracted from large biomedical event networks. Routine machine learning evaluation results, as well as manual evaluation results suggest that the problem is indeed learnable. This work won the Best Paper Award in The 5th International Symposium on Languages in Biology and Medicine (LBM 2013).

Wood-derived biomaterials for biomedical applications

Driven by the global trend in the sustainable economy development and environmental concerns, the exploring of plant-derived biomaterials or biocomposites for potential biomedical and/or pharmaceutical applications has received tremendous attention. Therefore, the work of this thesis is dedicated to high-value and high-efficiency utilization of plant-derived materials, with the focus on cellulose and hemicelluloses in the field of biomedical applications in a novel biorefinery concept. The residual cellulose of wood processing waste, sawdust, was converted into cellulose nanofibrils (CNFs) with tunable surface charge density and geometric size through 2,2,6,6-tetramethylpiperidinyloxy (TEMPO)-mediated oxidation and mechanical defibrillation. The sawdust-based CNFs and its resultant free-standing films showed comparable or even better mechanical properties than those from a commercial bleached kraft pulp at the same condition, demonstrating the feasibility of producing CNFs and films thereof with outstanding mechanical properties from birch sawdust by a process incorporated into a novel biorefinery platform recovering also polymeric hemicelluloses for other applications. Thus, it is providing an efficient route to upgrade sawdust waste to valuable products. The surface charge density and geometric size of the CNFs were found to play key roles in the stability of the CNF suspension, as well as the gelling properties, swelling behavior, mechanical stiffness, morphology and microscopic structural properties, and biocompatibility of CNF-based materials (i.e. films, hydrogels, and aerogels). The CNFs with tunable surface chemistry and geometric size was found promising applications as transparent and tough barrier materials or as reinforcing additive for production of biocomposites. The CNFs was also applied as structural matrices for the preparation of biocomposites possessing electrical conductivity and antimicrobial activity by in situ polymerization and coating of polypyrrole, and incorporation of silver nanoparticles, which make the material possible for potential wound healing application. The CNF-based matrices (films, hydrogels, and aerogels) with tunable structural and mechanical properties and biocompatibility were further prepared towards an application as 3D scaffolds in tissue engineering. The structural and mechanical strength of the CNF matrices could be tuned by controlling the charge density of the nanocellulose, as well as the pH and temperature values of the hydrogel formation conditions. Biological tests revealed that the CNF scaffolds could promote the survival and proliferation of tumor cells, and enhance the transfection of exogenous DNA into the cells, suggesting the usefulness of the CNF-based 3D matrices in supporting crucial cellular processes during cell growth and proliferation. The CNFs was applied as host materials to incorporate biomolecules for further biomedical application. For example, to investigate how the biocompatibility of a scaffold is influenced by its mechanical and structural properties, these properties of CNF-based composite matrices were controlled by incorporation of different hemicelluloses (O-acetyl galactoglucomanan (GGM), xyloglucan (XG), and xylan) into CNF hydrogel networks in different ratios and using two different approaches. The charge density of the CNFs, the incorporated hemicellulose type and amount, and the swelling time of the hydrogels were found to affect the pore structure, the mechanical strength, and thus the cells growth in the composite hydrogel scaffolds. The mechanical properties of the composite hydrogels were found to have an influence on the cell viability during the wound healing relevant 3T3 fibroblast cell culture. The thusprepared CNF composite hydrogels may work as promising scaffolds in wound healing application to provide supporting networks and to promote cells adhesion, growth, and proliferation.

Performance Analysis of IP based WSNs in real time systems

Wireless sensor networks (WSNs) are the key enablers of the internet of things (IoT) paradigm. Traditionally, sensor network research has been to be unlike the internet, motivated by power and device constraints. The IETF 6LoWPAN draft standard changes this, defining how IPv6 packets can be efficiently transmitted over IEEE 802.15.4 radio links. Due to this 6LoWPAN technology, low power, low cost micro- controllers can be connected to the internet forming what is known as the wireless embedded internet. Another IETF recommendation, CoAP allows these devices to communicate interactively over the internet. The integration of such tiny, ubiquitous electronic devices to the internet enables interesting real-time applications. This thesis work attempts to evaluate the performance of a stack consisting of CoAP and 6LoWPAN over the IEEE 802.15.4 radio link using the Contiki OS and Cooja simulator, along with the CoAP framework Californium (Cf). Ultimately, the implementation of this stack on real hardware is carried out using a raspberry pi as a border router with T-mote sky sensors as slip radios and CoAP servers relaying temperature and humidity data. The reliability of the stack was also demonstrated during scalability analysis conducted on the physical deployment. The interoperability is ensured by connecting the WSN to the global internet using different hardware platforms supported by Contiki and without the use of specialized gateways commonly found in non IP based networks. This work therefore developed and demonstrated a heterogeneous wireless sensor network stack, which is IP based and conducted performance analysis of the stack, both in terms of simulations and real hardware.