| 摘要 |
<p>A complete macromolecular structure (M) is constructed based upon two or more partial coordinate strings (W) representing the three-dimensional coordinates of the location of atoms (A) known to reside in some predetermined sequence at some predetermined region of the molecule. A library of known protein structures is searched to identify each structural fragment that has at least some predetermined subset of the atoms thereof with coordinates that correspond within predetermined limits imposed by a predetermined structural context mask to the coordinates of the atoms in each partial coordinate string (W). Each fragment identified by the search has assigned to it a context dependent probability quantifying the probability that the particular structural fragment corresponds to the partial coordinate string (W). Once all of the identified database fragments have been ranked in accordance with their context dependent probabilities, the coordinates of the atoms forming the structural fragment having the highest context dependent probability is substituted into the partial coordinate string (W), thereby defining an updated partial coordinate string (W). The resultant updated partial coordinate string (W) is tested in accordance with predetermined test parameters.</p> |
