Research Capabilities
Biosignature Integration for Inference of Biomarkers from Complex Systems
The project will result in a framework for describing data from one or multiple sources, diagnostic models of exposure or response for model systems based on data from multiple sources and inference of biomarkers and associated biological models relevant to the complex systems under evaluation.
PNNL is developing an approach to discovery-based biomarker identification using existing and new technologies at scales ranging from individual cells to ecosystems. The data integration to meet this goal requires collaboration among skill sets that include statistics, computer science, bioinformatics and biology. The first phase of this data integration effort uses a bioinformatics approach where Bayesian statistics support the injection of data-related information as prior probabilities in the form of expert opinion for making probabilistic inferences. The subsequent, more challenging phase of inferring biology uses a Bayesian methodology to allow inference of relationships between biomolecules and elucidation of biological activity unique to biological response. This data integration with the end goal of inferring system level biological activity from multiple technologies is the challenge.