Meta-models are statistical prediction models of the input-output relationship of dynamic models. These models can be analysed by a whole battery of data-analytic tools to facilitate the elucidation of a dynamic model's sensitivity to a parameter across a very wide range of parameter backgrounds and the epistasis relationships between parameters, the computational compaction of models, the comparison of models addressing overlappings sets of phenomena and for parameter estimation from empirical data. We are working hard to develop a battery of methods suited to construct and analyse multivariate metamodels of high-dimensional dynamic models characterized by high-dimensional parameter and state variable spaces. We are particularly focused on providing tools that can substantially improve the construction and validation of multiscale dynamic models with complex multivariate spatio-temporal phenotypes. In this context we are very much interested in making connections to research groups developing such multiscale models who may be interested in testing out the toolbox we are in the process of developing.