Multiple Alignment of Structures Using Center Of ProTeins

Protein structure comparison has long been under investigation by computational biologists in hopes for finding a better but not necessarily faster alternative to the more familiar multiple sequence alignment problem. The multiple alignment problem is more challenging than pairwise alignment even for sequences, and we resort to heuristics to find as best an approximation as possible, in polynomial time. Since officially, as of 2014, there are more protein structures in the Protein Data Bank[3] than seconds in a day, there is a constant need for both speed and precision when aligning more than two proteins. Multiple Structure Alignment (MStA) of protein structures can be categorized into four widely different approaches progressive alignment, core optimization, graph based, and pivot based. Mustang [13], msTali [23] , mulPBA [14], and CE-MC [7] use the progressive alignment approach that creates an alignment of alignments following a guide tree. While this approach does make sense, it suffers from the natural disadvantages of all progressive techniques. Methods from other approaches, [17] [29], often outperform the progressive ones, both in terms of speed and accuracy. A second way is to optimize a consensus structure, sometimes with several iterations, and report a common core of the input proteins. The idea is to find out a structurally conserved subset of residues among the proteins to gain some insight into their origin. However, such cores are mostly pseudo-structures that, although