Correlation and unification of computational and experimental findings

This part of the project will require statistical evaluation of the correspondence between available in vivo toxicity data, in vitro occurrence of biochemical markers, and in silico-modeled thermodynamic parameters for given chemicals. We will use proven statistical approaches such as including nearest-neighbor clustering, partition analysis and multivariate modeling to explore the descriptors in pairwise and multivariate fashions. These analyses will establish whether the anticipated and mechanistically rationalized relationships exist between the in silico parameters and the in vitro / in vivo data.