Intermediate sequences increase the detection of homology between sequences

J Mol Biol. 1997 Oct 17;273(1):349-54. doi: 10.1006/jmbi.1997.1288.

Abstract

Two homologous sequences, which have diverged beyond the point where their homology can be recognised by a simple direct comparison, can be related through a third sequence that is suitably intermediate between the two. High scores, for a sequence match between the first and third sequences and between the second and the third sequences, imply that the first and second sequences are related even though their own match score is low. We have tested the usefulness of this idea using a database that contains the sequences of 971 protein domains whose structures are known and whose residue identities with each other are some 40% or less (PDB40D). On the basis of sequence and structural information, 2143 pairs of these sequences are known to have an evolutionary relationship. FASTA, in an all-against-all comparison of the sequences in the database, detected 320 (15%) of these relationships as well as three false positive (i.e. 1% error rate). Using intermediate sequences found by FASTA matches of PDB40D sequences to those in the large non-redundant OWL database we could detect 550 evolutionary relationships with an error rate of 1%. This means the intermediate sequence procedure increases the ability to recognise the evolutionary relationships amongst the PDB40D sequences by 70%.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Amino Acid Sequence
  • Ascorbate Oxidase / chemistry
  • Bacterial Proteins / chemistry
  • Databases as Topic
  • Evolution, Molecular
  • Molecular Sequence Data
  • Plastocyanin / chemistry
  • Proteins / chemistry*
  • Sequence Homology, Amino Acid*
  • Software

Substances

  • Bacterial Proteins
  • Proteins
  • mauC protein, Methylobacterium extorquens
  • Plastocyanin
  • Ascorbate Oxidase