Optimal information propagation across scales in multi-scale integrative modeling

Determine the optimal strategy for propagating information across scales in multi-scale simulations used for integrative structural modeling of macromolecular assemblies, specifying how constraints and data should be exchanged between higher-resolution (atomic) and lower-resolution (coarse-grained) representations during sampling and refinement.

Background

The paper introduces optimizing model representations for integrative structural modeling, including incremental coarse-graining and Bayesian model selection via NestOR, to achieve efficient sampling and accurate data-to-model mapping. These approaches often yield non-uniform multi-scale representations where regions with more data are modeled at higher resolution, and others at lower resolution.

Within this context, the authors explicitly identify an open question on how best to propagate information across these scales during simulation. Resolving this question is important for maintaining consistency, accuracy, and efficiency in multi-scale integrative modeling workflows and is relevant to meta-modeling efforts that combine models at different scales.