OPTIMUM SUPERIMPOSITION OF PROTEIN STRUCTURES - AMBIGUITIES AND IMPLICATIONS

Background: Techniques for comparison and optimum superimposition of p rotein structures are indispensable tools, providing the basis for sta tistical analysis, modeling, prediction and classification of protein folds. Observed similarity of structures is frequently interpreted as an indication of evolutionary relatedness. A variety of advanced techn iques are available, but so far the important issue of uniqueness of s tructural superimposition has been largely neglected. We set out to in vestigate this issue by implementing an efficient algorithm for struct ure superimposition enabling routine searches for alternative alignmen ts. Results: The algorithm is based on optimum superimposition of stru ctures and dynamic programming. The implementation is tested and valid ated using published results. In particular, an automatic classificati on of all protein folds in a recent release of the protein data bank i s performed. The results obtained are closely related to published dat a. Surprisingly, for many protein pairs alternative alignments are obt ained. These alignments are indistinguishable in terms of number of eq uivalent residues and root mean square error of superimposition, but t he respective sets of equivalent residue pairs are completely distinct . Alternative alignments are observed for all protein architectures, i ncluding mixed alpha/beta folds. Conclusions: Superimposition of prote in folds is frequently ambiguous. This has several implications on the interpretation of structural similarity with respect to evolutionary relatedness and it restricts the range of applicability of superimpose d structures in statistical analysis. In particular, studies based on the implicit assumption that optimum superimposition of structures is unique are bound to be misleading.