《karlinandatlschul1990》.pdf

Pro. ,Vut!. Ac-(Id. Sei. USA Vol. 87. pp. 1264-ZZh8. March 1990 Evolution Methods for assessing the statistical significance of molecular sequence features by using general scoring schemes (sequence alignment/protein sequence features) S A M U E L KARL IN^ .4ND STEPHEN F. ALTSCHULSB ’Depanment of Mathematics. Stanford Univcrsitv. Stanford. CA 94305: and :National Center for Biotechnology Informa1ion. National Library of Medicine, .~ National Institutes of Health. Bethesda. MD 20844 Conrribured by Somuel Karlin. December 26, 1989 ABSTRACT An unusualpattern in a nucleicacid or pro- pendencies. For these models, theoretical results (distribu- or a region of strong similarity shared by two or for a variety tein sequence tional properties) have previously been obtained more sequences may have biological significance.It is therefore as the length of the longest run of of sequence statistics such desirable to knowwhether suchapatterncan have arisen agiven letter or pattern(allowing for afixednumber of simply by chance. To identify interesting sequence patterns, errors),the length of thelongestword(oligonucleotide, appropriate scoring values can be assigned to the individual peptide) in a sequence satisfying a prescribed relationship residues of a single sequenceor to sets of residues when several (e.g., r-fold repeat, dyad pairing), and counts andspacings of sequences are compared. For singl