Elucidation of Functional Networks from Toxicogenomics Gene Expression Data

DESCRIPTION (provided by applicant): In Phase I, we have developed a novel approach to the analysis of toxicogenomics data based on functional categorization and signature networks. The method enables identification of pathways and biological networks most different between gene expression response to drug action, and it can be applied for categorization of molecular profiles by toxicity categories. We have also functional profiling on a large dataset of hepatotoxic compounds from CEBS database. In Phase II, we propose to formalize the new approach in algorithms and build a semi-automated module for generation of signature networks specific for a toxicity type. We will also build a reference database of quantitative functional descriptors for toxic categories which can be used for categorizing of molecular profiles for novel compounds. We will develop a workflow for functional analysis of toxicogenomics datasets and a reporting protocol which sponsor companies can use for voluntary submission of toxicogenomics data to FDA and the reviewers for evaluation of the submitted data. Finally, we propose to build a novel integrated platform for functional data mining of toxicogenomics data and reporting, MetaTox. MetaTox will include the manually curated database of human biology created at GeneGo, a general datamining toolkit (networks, pathways, cellular processes, disease and toxicity categories), pre-processed toxicogenomics data from public domain, and the tools created in the scope of this project (signature networks module and the database of functional descriptors for toxicity categories). In Phase II, we propose to complete the set of tools and workflows for functional analysis of toxicogenomics data. This will include a novel module for generation and comparison of signature networks for different toxic categories, matrix of quantitative functional categories and a new toxicogenomics platform MetaTox