Adjunct Professor Igor Goryanin
goryanin at oist.jp
Human networks reconstruction
Our Human Metabolic reconstruction is being extended for a better understanding of biological processes in disease and healthy states.
The metabolic network is being integrated with the signal transduction network, gene regulatory network and possible protein-protein interaction networks from high-throughput experiments and de novo predictions based on developed Bioinformatics methods. The components in the integrated network (genes, proteins, metabolites) are linked with other information on human diseases and drugs, etc. Databases include KEGG, HMDB, OMIM, the Diseases Database, Genetic Association Database, PharmaGKB, MedlinePlus, eMedicine, STKE, GPDB, Nature PID.
The human metabolic network is being improved by increasing the number of annotated reactions, metabolites, and tissue distribution. Of particular interest is the need to integrate information about the possible sources of compounds in the human body and ways of their degradation as not all of them are biosynthesised by host-derived enzymes. The focus is on the drugs and signal metabolites (may be produced by human or another organism) related metabolic pathways as these metabolites often have important regulatory functions and can cause great perturbation of the state of the cell through complex signal transduction pathways.
The result will be a central one-stop service for network based biomedical analysis of human biological processes. A semi-automated pipeline is being developed for the reconstruction and validation of genome scale biological networks from databases and literature. We extract, standardize, represent and store the information in a structured way, and apply this knowledge for further in silico analysis. We are developing various network based methods to enable complex functional analysis .
The database allow export/import pathway models in the form of the Systems Biology Markup Language (www.sbml.org) and visualize pathway maps based on the Systems Biology Graphical Notation (www.sbgn.org), and other emerging community standards.
Novel computational approaches
The focus is on numerical techniques, novel methods and parallel software solutions for high-performance computers.
The software is being tested on specific models of signaling pathways with real experimental data.
New global sensitivity coefficients for analysis of qualitative semi-qualitative and quantitative behavior of systems with large numbers of non-identifiable parameters.
Novel population-based optimization and game theory methods for parameter identification and reconstruction of evolving system properties
Novel methods of multi- and mega-variate data analysis to help in setting boundaries in parameter space for better convergence in multi-objective optimization.