Fast computer search for similar DNA sequences.

An extremely fast method of searching a nucleic acid sequence database against a probe sequence is described. The method is based on the detection of deviation from expected number and deviation from random spatial distribution of sub-sequences which are unique within a sequence, and shared between that sequence and the probe. On an IBM 3081 computer, total search of an encoded form of the EMBL nucleic acid sequence database with a 1 kbase probe sequence is completed in a few seconds. Previous best methods for a similar task required a few minutes.