The structural analysis of genomic DNA sequences is discussed in the framework of the spectral approach, which is sufficiently universal due to the reciprocal correspondence and mutual complementarity of Fourier transform length scales. The spectral characteristics of random sequences of the same nucleotide composition possess the property of self-averaging for relatively short sequences of length M \geqslant 100-300. Comparison with the characteristics of random sequences determines the statistical significance of the structural features observed. Apart from traditional applications to the search for hidden periodicities, spectral methods are also efficient in studying mutual correlations in DNA sequences. By combining spectra for structure factors and correlation functions, not only integral correlations can be estimated but also their origin identified. Using the structural spectral entropy approach, the regularity of a sequence can be quantitatively assessed. A brief introduction to the problem is also presented and other major methods of DNA sequence analysis described.

PACS: 02.50.−r, 05.10.−a, 87.10.+e, 87.14.Gg (all) DOI: 10.1070/PU2000v043n01ABEH000611 URL: https://ufn.ru/en/articles/2000/1/c/ 000086055100003 Citation: Lobzin V V, Chechetkin V R "Order and correlations in genomic DNA sequences. The spectral approach" Phys. Usp. 43 55–78 (2000) BibTexBibNote ® (generic)BibNote ® (RIS)MedlineRefWorks Оригинал: Лобзин В В, Чечеткин В Р «Порядок и корреляции в геномных последовательностях ДНК. Спектральный подход» УФН 170 57–81 (2000); DOI: 10.3367/UFNr.0170.200001c.0057