Electron microscopy has become an important tool in structural biology. Many proteins that cannot be made to crystallize in a 3D crystal suitable for X-ray crystallography can nevertheless be studied by electron microscopy, either as single molecules or in the form of 2D crystals. However, the resulting structures have a much lower resolution. With rare exceptions, the electron density maps obtained by electron microscopy are not sufficient to identify the structure at atomic resolution.

In some situations, an atomic-level structure of the same protein in another conformation is available. DensityFit combines this information with the low-resolution structure by deforming the known atomic structure to make it fit into the low-resolution electron density map. The flexibility of the protein is described by an elastic network model, whose low-energy normal modes define the allowed deformations. This procedure guarantees that the integrity of the protein is preserved during the deformation.

Compared to other similar algorithms that have been published, the one implemented in DensityFit has the advantage of having been tested on real experimental data for a mid-size protein, Ca-ATPase. The algorithm and its application to Ca-ATPase have been described in

  • K Hinsen, N Reuter, J Navaza, D L Stokes, and J J Lacapere
    Normal mode based fitting of atomic structure into electron density maps: application to SR Ca-ATPase
    Biophys. J. 88:818–827 (2005)

The underlying normal mode techniques and the harmonic model being used for the deformation of the protein are described in

For downloading DensityFit, and for installation instructions, see the source code repository.