The force field and Monte Carte sampling method of our recently developed r
educed model of proteins is described. Recent applications of the models in
clude nb initio structure prediction for small globular proteins, modeling
of protein structure based on distantly homologous (or analogous) structura
l templates, assembly of protein structure from sparse experimental data, a
nd computational studies of protein folding dynamics and thermodynamics. Th
e newest application, described in this paper, enables the prediction of lo
w-to-moderate resolution coordinates of the parts of protein structure that
are missed in incomplete PDB files.