The effect of sequence correlation on bubble statistics in double-stranded DNA.

DNA exists stably in the double-stranded structure at physiological temperatures, but base pairs are observed to unbind locally, giving way to bubbles (i.e., locally denatured states) due to thermal fluctuation. In this study, we consider the effect of sequence on the bubble statistics. On the basis of the Edwards equation description [W. Sung and J.-H. Jeons, Phys. Rev. E 69, 031902 (2004) ], we develop a stochastic model incorporating the sequence randomness as a dichotomic noise, where the bubble and its size are identified as a returning random walk and its first passage time, respectively. By simulating the model Langevin equation, we obtain the bubble size distribution and show how it is affected by the sequence correlation. We find that the bubble size distribution of DNA with finite sequence correlation deviates from the Poland-Scheraga-type distribution. In particular, the formation of large bubbles is dramatically enhanced as sequence correlation length gets longer.