Analysis of sequences of twist angles in DNA double helix.

By assuming that the realistic DNA chains are random sequence of bases and using the Tung-Harvey formula for the prediction of twist angles, it is shown that the mean value of the sequence of twist angles is almost sequence-independent. In general the variance for the A, T-rich sequence is larger than that of G, C-rich sequence. There exists an upper bound for the variance of all possible sequences, i.e., the variance is not greater than 27 deg2. It is pointed out that the large conformational deviation from ideal DNA is an important factor for the recognition of DNA with protein/enzyme.