@phdthesis{Cecil2012, author = {Cecil, Alexander [geb. Schmid]}, title = {Metabolische Netzwerkanalysen f{\"u}r den Weg von xenobiotischen zu vertr{\"a}glichen antibiotischen Substanzen}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-71866}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2012}, abstract = {Durch das Auftreten neuer St{\"a}mme resistenter Krankheitserreger ist die Suche nach neuartigen Wirkstoffen gegen diese, sich st{\"a}ndig weiter ausbreitende Bedrohung, dringend notwendig. Der interdisziplin{\"a}re Sonderforschungsbereich 630 der Universit{\"a}t W{\"u}rzburg stellt sich dieser Aufgabe, indem hier neuartige Xenobiotika synthetisiert und auf ihre Wirksamkeit getestet werden. Die hier vorgelegte Dissertation f{\"u}gt sich hierbei nahtlos in die verschiedenen Fachbereiche des SFB630 ein: Sie stellt eine Schnittstelle zwischen Synthese und Analyse der Effekte der im Rahmen des SFB630 synthetisierten Isochinolinalkaloid-Derivaten. Mit den hier angewandten bioinformatischen Methoden wurden zun{\"a}chst die wichtigsten Stoffwechselwege von S. epidermidis R62A, S. aureus USA300 und menschlicher Zellen in sogenannten metabolischen Netzwerkmodellen nachgestellt. Basierend auf diesen Modellen konnten Enzymaktivit{\"a}ten f{\"u}r verschiedene Szenarien an zugesetzten Xenobiotika berechnet werden. Die hierf{\"u}r ben{\"o}tigten Daten wurden direkt aus Genexpressionsanalysen gewonnen. Die Validierung dieser Methode erfolgte durch Metabolommessungen. Hierf{\"u}r wurde S. aureus USA300 mit verschiedenen Konzentrationen von IQ-143 behandelt und gem{\"a}ß dem in dieser Dissertation vorgelegten Ernteprotokoll aufgearbeitet. Die Ergebnisse hieraus lassen darauf schließen, dass IQ-143 starke Effekte auf den Komplex 1 der Atmungskette aus{\"u}bt - diese Resultate decken sich mit denen der metabolischen Netzwerkanalyse. F{\"u}r den Wirkstoff IQ-238 ergaben sich trotz der strukturellen {\"A}hnlichkeiten zu IQ-143 deutlich verschiedene Wirkeffekte: Dieser Stoff verursacht einen direkten Abfall der Enzymaktivit{\"a}ten in der Glykolyse. Dadurch konnte eine unspezifische Toxizit{\"a}t dieser Stoffe basierend auf ihrer chemischen Struktur ausgeschlossen werden. Weiterhin konnten die bereits f{\"u}r IQ-143 und IQ-238 auf Bakterien angewandten Methoden erfolgreich zur Modellierung der Effekte von Methylenblau auf verschiedene resistente St{\"a}mme von P. falciparum 3D7 angewandt werden. Dadurch konnte gezeigt werden, dass Methylenblau in einer Kombination mit anderen Pr{\"a}paraten gegen diesen Parasiten zum einen die Wirkung des Prim{\"a}rpr{\"a}parates verst{\"a}rkt, zum anderen aber auch in gewissem Maße vorhandene Resistenzen gegen das Prim{\"a}rpr{\"a}parat zu verringern vermag. Somit konnte durch die vorgelegte Arbeit eine Pipeline zur Identifizierung der metabolischen Effekte verschiedener Wirkstoffe auf unterschiedliche Krankheitserreger erstellt werden. Diese Pipeline kann jederzeit auf andere Organismen ausgeweitet werden und stellt somit einen wichtigen Ansatz um Netzwerkeffekte verschiedener, potentieller Medikamente aufzukl{\"a}ren.}, subject = {Stoffwechsel}, language = {de} } @phdthesis{Nilla2012, author = {Nilla, Jaya Santosh Chakravarthy}, title = {An Integrated Knowledgebase and Network Analysis Applied on Platelets and Other Cell Types}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-85730}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2012}, abstract = {Systems biology looks for emergent system effects from large scale assemblies of molecules and data, for instance in the human platelets. However, the computational efforts in all steps before such insights are possible can hardly be under estimated. In practice this involves numerous programming tasks, the establishment of new database systems but as well their maintenance, curation and data validation. Furthermore, network insights are only possible if strong algorithms decipher the interactions, decoding the hidden system effects. This thesis and my work are all about these challenges. To answer this requirement, an integrated platelet network, PlateletWeb, was assembled from different sources and further analyzed for signaling in a systems biological manner including multilevel data integration and visualization. PlateletWeb is an integrated network database and was established by combining the data from recent platelet proteome and transcriptome (SAGE) studies. The information on protein-protein interactions and kinase-substrate relationships extracted from bioinformatical databases as well as published literature were added to this resource. Moreover, the mass spectrometry-based platelet phosphoproteome was combined with site-specific phosphorylation/ dephosphorylation information and then enhanced with data from Phosphosite and complemented by bioinformatical sequence analysis for site-specific kinase predictions. The number of catalogued platelet proteins was increased by over 80\% as compared to the previous version. The integration of annotations on kinases, protein domains, transmembrane regions, Gene Ontology, disease associations and drug targets provides ample functional tools for platelet signaling analysis. The PlateletWeb resource provides a novel systems biological workbench for the analysis of platelet signaling in the functional context of protein networks. By comprehensive exploration, over 15000 phosphorylation sites were found, out of which 2500 have the corresponding kinase associations. The network motifs were also investigated in this anucleate cell and characterize signaling modules based on integrated information on phosphorylation and protein-protein interactions. Furthermore, many algorithmic approaches have been introduced, including an exact approach (heinz) based on integer linear programming. At the same time, the concept of semantic similarities between two genes using Gene Ontology (GO) annotations has become an important basis for many analytical approaches in bioinformatics. Assuming that a higher number of semantically similar gene functional annotations reflect biologically more relevant interactions, an edge score was devised for functional network analysis. Bringing these two approaches together, the edge score, based on the GO similarity, and the node score, based on the expression of the proteins in the analyzed cell type (e.g. data from proteomic studies), the functional module as a maximum-scoring sub network in large protein-protein interaction networks was identified. This method was applied to various proteome datasets (different types of blood cells, embryonic stem cells) to identify protein modules that functionally characterize the respective cell type. This scalable method allows a smooth integration of data from various sources and retrieves biologically relevant signaling modules.}, subject = {Systembiologie}, language = {en} } @phdthesis{Boyanova2012, author = {Boyanova, Desislava Veselinova}, title = {Systems biological analysis of the platelet proteome and applications of functional module search in proteome networks}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-72165}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2012}, abstract = {Recent development of proteomic approaches and generation of large-scale proteomic datasets calls for new methods for biological interpretation of the obtained results. Systems biological approaches such as integrated network analysis and functional module search have become an essential part of proteomic investigation. Proteomics is especially applied in anucleate cells such as platelets. The underlying molecular mechanisms of platelet activation and their pharmacological modulation are of immense importance for clinical research. Advances in platelet proteomics have provided a large amount of proteomic data, which has not yet been comprehensively investigated in a systems biological perspective. To this end, I assembled platelet specific data from proteomic and transcriptomic studies by detailed manual curation and worked on the generation of a comprehensive human platelet repository for systems biological analysis of platelets in the functional context of integrated networks (PlateletWeb) (http:/PlateletWeb.bioapps.biozentrum.uni-wuerzburg.de). I also added platelet-specific experimentally validated phosphorylation data and generated kinase predictions for 80\% of the newly identified platelet phosphosites. The combination of drug, disease and pathway information with phosphorylation and interaction data makes this database the first integrative platelet platform available for platelet research. PlateletWeb contains more than 5000 platelet proteins, which can also be analyzed and visualized in a network context, allowing identification of all major signaling modules involved in platelet activation and inhibition. Using the wealth of integrated data I performed a series of platelet-specific analyses regarding the platelet proteome, pathways, drug targets and novel platelet phosphorylation events involved in crucial signaling events. I analyzed the statistical enrichment of known pathways for platelet proteins and identified endocytosis as a highly represented pathway in platelets. Further results revealed that highly connected platelet proteins are more often targeted by drugs. Using integrated network analysis offered by PlateletWeb, I analyzed the crucial activation signaling pathway of adenosine diphosphate (ADP), visualizing how the signal flow from receptors to effectors is maintained. My work on integrin inside-out signaling was also based on the integrated network approach and examined new platelet-specific phosphorylation sites and their regulation using kinase predictions. I generated hypothesis on integrin signaling, by investigating the regulation of Ser269 phosphorylation site on the docking protein 1 (DOK1). This phosphorylation site may influence the inhibiting effect of DOK1 on integrin a2bb3. Extending the integrated network approach to further cell lines, I used the assembled human interactome information for the analysis of functional modules in cellular networks. The investigation was performed with a previously developed module detection algorithm, which finds maximum-scoring subgraphs in transcriptomic datasets by using assigned values to the network nodes. We extended the algorithm to qualitative proteomic datasets and enhanced the module search by adding functional information to the network edges to concentrate the solution onto modules with high functional similarity. I performed a series of analyses to validate its performance in small-sized (virus-infected gastric cells) and medium-sized networks (human lymphocytes). In both cases the algorithm extracted characteristic modules of sample proteins with high functional similarity. The functional module search is especially useful in site-specific phosphoproteomic datasets, where kinase regulation of the detected sites is often sparse or lacking. Therefore, I used the module detection algorithm in quantitative phosphoproteomic datasets. In a platelet phosphorylation dataset, I presented a pipeline for network analysis of detected phosphorylation sites. In a second approach, the functional module detecting algorithm was used on a phosphoproteome network of human embryonic stem cells, in which nodes represented the maximally changing phosphorylation sites in the experiment. Additional kinases from the human phosphoproteome in PlateletWeb were included to the network to investigate the regulation of the signal flow. Results indicated important phosphorylation sites and their upstream kinases and explained changes observed in embryonic stem cells during differentiation. This work presents novel approaches for integrated network analysis in cells and introduces for the first time a systematic biological investigation of the human platelet proteome based on the platelet-specific knowledge base PlateletWeb. The extended methods for optimized functional module detection offer an invaluable tool for exploring proteomic datasets and covering gaps in complex large-scale data analysis. By combining exact module detection approaches with functional information data between interacting proteins, characteristic functional modules with high functional resemblance can be extracted from complex datasets, thereby focusing on important changes in the observed networks.}, subject = {Netzwerkanalyse}, language = {en} }