The vast majority of sequence analysis programs available to date are designed to allow users to detect contiguous sequence fragments that occur more frequently than other contiguous motifs. However, substantial experimental evidence indicates that specific functions of nucleic acids fragments are rarely associated with single, simple sequence motifs. The patterns responsible for biological function are most often non-contiguous and consist of several short oligonucleotides separated by gaps of variable length and composition. In view of the ongoing large sequencing projects reliable procedures for physical mapping of DNA functional domains will be invaluable. To make reliable maps one needs reliable criteria that distinguish between sequences involved in different functions. Non-contiguous pattern analysis already has provided a substantial number of criteria for distinction between protein coding genes, introns, and illegitimate recombination sites. It is believed that by using this methodology various kinds of regulatory regions within repetitive DNA sequences will be mapped unambiguously.
