A sequence-comparison statistic and algorithm

摘要: SUMMARY A statistic is proposed for the comparison of ordered sequences objects when aim to slot together sequences. An algorithm suggested attempting find best pairwise fit. fair number problems in earth and life sciences can be regarded terms comparing objects. Some these involve direct matching identical objects; example, could amino acids a protein sequence. Ways approaching this problem have been by Needleman & Wunsch (1970), Gibbs MacIntyre Sackin (1971). In many situations, however, one sequence cannot identified with other sequence, but merely resemble them greater or lesser degree. For pollen stratigraphical analyses, would correspond different levels within cores, at each level which there was count grains types. botanical studies along transects, locations we took measurements. These examples common measurement values certain attributes specified single variable: 'time-depth' stratigraphy 'distance' transect. both above, it interest know area object resembles, also consider overall resemblances We assume that define measure similarity between all pairs fits satisfactorily pair if its high. seek together, as shown, Table 1, so satisfactory local fits, subject constraints imposed orderings.