WP12 Predictive comparisons
Currently, predictive comparisons in liver and heart toxicity modelling are hardly available due to the lack of benchmark data. Although initial models and frameworks exist, they target rather heterogeneous aspects and have not yet been studied in acomparable fashion. Existing models are available from literature or, for models that are compliant with exchange formats (such as SBML or BioPAX) in public repositories or through dedicated modelling software. A very recent review by the US FDA describe advances in bioinformatics, computersciences, next-generation sequencing, and systems biology that create the opportunity for integrated systems pharmacology-based prediction of drug safety declares that success of this effort is based on the assumption that a systems network that consists of druginduced perturbations of physiological functions can be characterized. Therefore, this network should span the hierarchy of biological organization, from gene to mRNA to protein to intracellular organelle to cell to organ to organism. It should be populated with data from each of these levels of biological organization. The network should contain interrelated genomics, transcriptomics, and metabolomics data, as well as organ and physiological functional data that are derived from the universe of information that describes and analyzes drug toxicity. In compliance with this, HeCaToS will conduct an iterative cycle of experimentation, predictive computer modelling, validation and refinement of computer modelling with the ultimate purpose to predict toxic AOPs in two essential domains of hepato- and cardiotoxicity.
The goal of workpackage 12 is to specify, define, implement and conduct use cases in the context of liver and heart toxicity that connect the computational modelling work packages (SP1) with the data generation (SP2) and analysis (SP3) work packages based on methods developed in SP3.
Use cases will focus on relevant problems of modern toxicogenomics research such as
- in-depth characterization of compound action,
- construction of Adverse Outcome Pathawys (AOPs),
- development of prediction models.
The conduction of use cases comprises the following steps:
- specification of the hypothesis and focus
- collection and generation of benchmark data
- schema for computational analysis
- validation of prediction results
- description and documentation of complete study.