A fast, sensitive pattern-matching approach for protein sequences

Pattern-matching algorithms are a powerful tool for findingsimilarities and relationships among the steadily growing amountof known protein sequences. We present a fast, sensitive pattern-matchingalgorithm that describes a pattern by its physico-chemical propertiesrather than by occurrence ofamino acids, using a fast, dynamicprogramming algorithm. Selected examples will demonstrate applicationsand advantages of our approach.