A novel method for multiple alignment of sequences with repeated and shuffled elements

Academic Article


  • We describe ABA (A-Bruijn alignment), a new method for multiple alignment of biological sequences. The major difference between ABA and existing multiple alignment methods is that ABA represents an alignment as a directed graph, possibly containing cycles. This representation provides more flexibility than does a traditional alignment matrix or the recently introduced partial order alignment (POA) graph by allowing a larger class of evolutionary relationships between the aligned sequences. Our graph representation is particularly well-suited to the alignment of protein sequences with shuffled and/or repeated domain structure, and allows one to construct multiple alignments of proteins containing (1) domains that are not present in all proteins, (2) domains that are present in different orders in different proteins, and (3) domains that are present in multiple copies in some proteins. In addition, ABA is useful in the alignment of genomic sequences that contain duplications and inversions. We provide several examples illustrating the applications of ABA. © 2004 by Cold Spring Harbor Laboratory Press.
  • Authors

    Digital Object Identifier (doi)

    Author List

  • Raphael B; Zhi D; Tang H; Pevzner P
  • Start Page

  • 2336
  • End Page

  • 2346
  • Volume

  • 14
  • Issue

  • 11