This paper describes an algorithm for retrieving sequences from genomic dat
abases. The algorithm presented offers a higher sensitivity by utilizing an
absolute metric for scoring the similarity query and the database sequence
s. Such a metric is an improvement over the existing methodology that utili
zes a string-edit or Levenshtein distance. A two-phase sequence retrieval s
ystem is built using this methodology and the results presented demonstrate
its advantages. (C) 2001 Elsevier Science B.V. All rights reserved.