Protein structure comparison: The quest for a meaningful measure of similarity

To a biochemist there is perhaps nothing so elusive as the three dimensional structure of a favourite protein. As such the development of a fast and reliable method to computationally determine protein structure is something of a holy grail, and presents one of the major challenges to science. In the assessment of the reliability of a given prediction technique, it is necessary to quantitatively compare the differences between a predicted output, and the desired one. Here we examine the shortcomings of some traditional structure comparison techniques, as well as presenting the challenges brought by new methods.