Modeling of complex protein structures

The solution structure of simple and complex proteins can be modeled efficaciously, starting either from solution scattering data or from the 3D structures derived from crystallographic work or from electron microscopy. Best fitting low-resolution bead models of proteins in solution can be retrieved successfully from the experimental scattering profiles by a genetic algorithm, in addition to rather conventional trials based on parametric data. The atomic coordinates of proteins, stored in several databases, can be exploited to model static, high-resolution images of anhydrous proteins. Additional use of appropriate surface calculation and hydration approaches allows more realistic hydrated protein models to be derived. Three dimensional reconstructions from electron microscopy represent very accurate images of proteins. However, they also require adequate consideration of the hydration contributions implied, for example, by a sophisticated analysis of the voxel density distribution, to simulate biological relevant entities. In all cases, the hydrated protein models yield much more accurate predictions of structural and hydrodynamic data than the anhydrous models.