MUSTA - A general, efficient, automated method for multiple structure alignment and detection of common motifs: Application to proteins

N. Leibowitz, R. Nussinov*, H. J. Wolfson

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Here we present an algorithm designed to carry out multiple structure alignment and to detect recurring substructural motifs. So far we have implemented it for comparison of protein structures. However, this general method is applicable to comparisons of RNA structures and to detection of a pharmacophore in a series of drug molecules. Further, its sequence order independence permits its application to detection of motifs on protein surfaces, interfaces, and binding/active sites. While there are many methods designed to carry out pairwise structure comparisons, there are only a handful geared toward the multiple structure alignment task. Most of these tackle multiple structure comparison as a collection of pairwise structure comparison tasks. The multiple structural alignment algorithm presented here automatically finds the largest common substructure (core) of atoms that appears in all the molecules in the ensemble. The detection of the core and the structural alignment are done simultaneously. The algorithm begins by finding small substructures that are common to all the proteins in the ensemble. One of the molecules is considered the reference; the others are the source molecules. The small substructures are stored in special arrays termed combinatorial buckets, which define sets of multistructural alignments from the source molecules that coincide with the same small set of reference atoms (Cα-atoms here). These substructures are initial small fragments that have congruent copies in each of the proteins. The substructures are extended, through the processing of the combinatorial buckets, by clustering the superpositions (transformations). The method is very efficient.

Original languageEnglish
Pages (from-to)93-121
Number of pages29
JournalJournal of Computational Biology
Issue number2
StatePublished - 2001


  • Geometric hashing
  • Invariants
  • Multiple structure alignment
  • Multiple structure comparison
  • Structural core


Dive into the research topics of 'MUSTA - A general, efficient, automated method for multiple structure alignment and detection of common motifs: Application to proteins'. Together they form a unique fingerprint.

Cite this