Detection of Hidden Repeats in Primary Structures of Protein Macromolecules Based on Spectral Approach
摘要
This work shows that detection of imperfect repeats at the level of primary sequences is possible within the framework of the spectral method. The study of primary sequences of protein macromolecules reveals nonobvious patterns that are obtained at the limit of the resolution of the spectral method. The detected hidden repeats have 25% matches in the Hamming metric, compared to relatively easily detected imprecise repeats, which have more than 50% matches. The validity of the results follows from the consideration of the dot matrix and the density of the correlation integral obtained by the spectral method. It was found that the spectral approach, which does not use predefined statistical models trained on bioinformatics material, nevertheless turns out to be a very sensitive tool for detecting and studying the structure of repeats at the level of the primary sequence. It has been shown that hidden repeats may be directly related to repeats in the spatial structure of a protein macromolecule.