SEQUENCE COMPOSITIONAL COMPLEXITY OF DNA THROUGH AN ENTROPIC SEGMENTATION METHOD

A new complexity measure, based on the entropic segmentation of DNA se quences into compositionally homogeneous domains, is proposed, Sequenc e compositional complexity (SCC) deals directly with the complex heter ogeneity in nonstationary DNA sequences, The plot of SCC as a function of significance level provides a profile of sequence structure at dif ferent length scales, SCC is found to be higher in sequences with long -range correlation than those without, and higher in noncoding sequenc es than coding sequences. Furthermore, a general agrement is found bet ween the SCC of the DNA sequence, on one hand, and the biological comp lexity of the organism, on the other, attributable to an increasingly complex organization of noncoding DNA over the course of evolution.