Statistical correlation of nucleotides in a DNA sequence

Abstract

We review methods in the study of nucleotide correlation in DNA sequence, and demonstrate two basic properties of the correlation through statistical analysis, namely, the short-range dominance of nucleotide correlation in most DNA sequences and the coarse-grained evolutionary dependence of the short-range correlation in coding sequences. A corresponding evolutionary mechanism is suggested. By the use of spectral analysis a large inhomogeneity in long-range base correlations for different sequences is indicated. Some results on three-dimensional DNA walks are reported. The linguistic differences between coding and noncoding sequences are also indicated.

Information decomposition method to analyze symbolical sequences

“…The results obtained using this method demonstrated that a large number of known genetic texts contain sequences with latent periodicity of various lengths and various types. It is in agreement with the results of investigation of DNA sequences by integral mathematical methods and finding the long-range correlations (LRC) in DNA sequences [55][56][57][58][59][60][61][62][63][64][65][66][67][68]. a -Alpha platelet-derived growth factor receptor precursor [52] from Rattus norvegicus (1088 amino acids, PGDS_RAT in Swiss-prot).…”

Time series model based on global structure of complete genome

“…The DNA walk defined in [5] is that the walker steps "up" if a pyrimidine (C or T ) occurs at position i along the DNA chain, while the walker steps "down" if a purine (A or G) occurs at position i. Stanley and coworkers [5] discovered there exists long-range correlation in noncoding DNA sequences while the coding sequences correspond to regular random walk. But if one considers more details by distinguishing C from T in pyrimidine, and A from G in purine (such as two or three dimensional DNA walk model [1] and maps given in [14]), then the presence of base correlation has been found even in coding region. However, DNA sequences are more complicated than those these types of analysis can describe.…”

“…A great deal of information concerning origin of life, evolution of species, development of individuals, and expression and regulation of genes, exist in these sequences [1] . In the past decade or so there has been an enormous interest in unravelling the mysteries of DNA.…”

Evolutionary direction of processed pseudogenes

“…Notably, our analysis indicated that, in pseudogenes, the increase in mutual information is not caused by a minor proportion of repeats (data not shown). In addition, we point out that, although base correlation (here, the mutual information) in coding sequences is likely to become stronger under certain kinds of selection and become weaker due to random mutations (Luo et al, 1998), this is not the case for pseudogenes. If the majority of the mutations in pseudogenes are assumed to occur either randomly (i.e., without selection of any particular type of mutations that have selective advantage), or under some kind of selection at population level, as will be discussed later on, the above described intrinsic mutational bias of DNA can lead to increased bias in the dinucleotide composition of the pseudogenes.…”

“…Mutual information has also been described as base correlation (Luo et al, 1998). Generalized mutual information for two nucleotides with a distance of k bp (k=0, 1, 2,…) can be defined as…”