We describe how to build protein models from structural templates. Methods to identify structural similarities between proteins in cases of significant, moderate to low, or virtually absent sequence similarity are discussed. The detection and evaluation of structural relationships is emphasized as a central aspect of protein modeling, distinct from the more technical aspects of model building. Computational techniques to generate and complement comparative protein models are also reviewed. Two examples, P-selectin and gp39, are presented to illustrate the derivation of protein model structures and their use in experimental studies.Keywords: protein modeling; protein structure; sequence similarity; sequence-structure compatibility; structural similarityThe gap between the number of available amino acid sequences and three-dimensional structures of proteins elucidated by crystallography or NMR techniques is expanding rapidly. The rate of sequence determination is at least 50-fold higher than the rate of structure determination (Bowie et al., 1991). It is therefore not surprising to note an increasing interest in predictive methods to derive three-dimensional protein models (Thornton et al., 1991;Fetrow & Bryant, 1993;Rost et al., 1993). We will review current protein modeling techniques, with a focus on knowledge-based methods (Blundell et al., 1987;Greer, 1991). Central to the review will be the question of how meaningful template structures for protein modeling can be identified and used for model building. Furthermore, we will report on two recent examples of knowledge-based model building carried out in our laboratory, P-selectin and gp39, the human ligand for CD40, and will discuss the role of these models for the rationalization and design of experiments.
Homology versus similarityKnowledge-based model building is often called "modeling by homology." Such modeling techniques start from ~~ ~ ~~