Matching relational patterns in nucleic acid sequences

Abstract

We describe a program that efficiently searches sequence data banks for complex patterns where sites are linked by common relations such as identity, complementarity or span. Its algorithm is closer to those of automatic demonstration than to the finite state machines used in fast pattern matching. The repertory of relations can be enriched at will without rewriting the core of the program. The program is written in Pascal-ISO and runs on a microcomputer.