Biological knowledge discovery through mining multiple sources of high-throughput data

摘要: As we are moving into the post-genomic era, various high-throughput experimental techniques have been developed to characterize biological systems at genome scale. The data becoming fundamentally important resources shed new insights on system-level understanding of ‘ organization’ and ‘dynamics’ molecules (i.e. genes, proteins), relationships between them, interaction cascades, pathways, modules networks regulation, co-expression metabolism). This dissertation focuses developing computational tools facilitate process translating ever-growing volumes significant knowledge protein functions, pathways modules. Although provide a global picture about underlying mechanisms, details often noisy, hence integration heterogeneous that cellular from different aspects gene expression protein-protein interactions) can lead comprehensive coherent discoveries insights. We Bayesian probability framework predict function for unannotated proteins in yeast through integrating binary data, complex microarray data. also extended infer pathway an automated systematical fashion. Besides bottom-up approaches functions applied top-down model network, is, started architecture network identify functional k-core algorithm decompose networks, which provides strong support modularity principles networks' structure function. Dynamic complexes identified by clustering constructed multiple sources shedding organization dynamics living cell. We proposed consensus approach combining In future, with explosion quantity diversity it is vital develop methodologies innovative bioinformatics explore iterative