"Structural databases of protein families and their use in fold recognition and comparative modelling"
Structure-based alignment of sequences of related proteins is key to understanding the structure and function of protein families. We have developed databases of structurally aligned homologous families (HOMSTRAD; http://www-cryst.bioc.cam.ac.uk/~homstrad/) and superfamilies (CAMPASS; http://www-cryst.bioc.cam.ac.uk/~campass/). In these databases, structure-based alignments are analyzed and formatted with the program JOY (http://www-cryst.bioc.cam.ac.uk/~joy/). The alignment databases and JOY proved to be particularly useful in the following two areas: 1) detection of distant homologues, where conventional sequence comparison methods fail to recognize such relationships, and 2) comparative modelling to gain insight into the 3D structure and function of a member of a homologous family. Successful applications in each of these areas will be described. The key idea is to identify conserved amino acids in their specific local environments, and this is more systematically exploited in our new fold recognition software, which can be used for large-scale genome analyses.