Multiple sequence alignment.

摘要: A method has been developed for aligning segments of several sequences at once. The number search steps depends only polynomially on the sequences, instead exponentially, because most alignments are rejected without being evaluated explicitly. data structure herein called "heap" facilitates this process. For a set n sequence segments, overall similarity is taken to be sum all constituent segment pair similarities, which in turn sums corresponding residue scores from Table. statistical models that test significance make it possible group objectively, even when or interrelationships weak. These tests very sensitive, while remaining quite conservative, and discourage addition "misfit" an existing set. new techniques applied five DNA-binding proteins, three enzymes employ coenzyme FAD, control alignment previously proposed proteins basis structural comparisons inspection supported dramatically, highly significant found FAD-binding proteins.