Recently available for ADME/Tox research, new """"""""genomics"""""""" technologies allowed to exponentially increase the throughput of pre-clinical toxicity testing. However, the functional analysis of new data lags behind, with mainly mechanistic clustering statistical procedures. We propose a novel approach of identification and analysis of functional networks and modules implicated in toxicity in human and experimental models. In Phase I we will use our system's biology data mining suit MetaCore to analyze publicly available toxicogenomics microarray expression data and reconstruct signaling, regulatory and metabolic networks specifically implicated in different types of hepatotoxicity in rat. The networks will be adjusted to reflect the difference in protein-protein interactions between human and rat and analyzed using multiple network comparison algorithms within our suit MetaCore. The resulted """"""""signature networks"""""""" characteristic for specific toxicities, will be tested in genome-wide microarray experiments in collaboration with the University of Wisconsin (EDGE ? 2 database). The project will result in a unique toolkit for rational analysis of gene expression, proteomics and metabolomics data widely needed in toxicogenomics research, and in series of """"""""signature networks"""""""" applicable for early detection and sub-categorization of hepatotoxicity. The technology allows integration and analysis of different types of experimental data on the same backbone of functional networks. The network analysis toolkit (algorithms, database, visualization tools) will be integrated in our ADME/Tox suit MetaDrug and offered to phramaceutical companies and academic institutions including NIEHS and FDA.
