Search of periodicities in primary structure of biopolymers: A general fourier approach

Abstract

We discuss a new convenient way to study periodical patterns in primary structures of biopolymers which appeared recently. For the sequence of a biopolymer the symbolic correlation function is constructed, which, is used as a digital sequence thus allowing us to perform a Fourier transform. Another fruitful technical improvement is the closing of the sequence in the ring with further scanning of the ring length, which allows the study of periods of the order of the sequence length. This approach makes it possible to take into account any scores describing similarity between symbols and to compare results obtained using different Fourier-like and correlation matrix techniques. An algorithm to compute Fourier spectrum power allows detection of vague periods in sequences containing strong repeats. A PASCAL program, SYMFOUR, has been written and tested on both sequences with periodical patterns, already reported, and sequences and other sites interesting from a biological point of view. © 1996, Oxford University Press.