We present a method for calculating predicted locations and extents of stress-induced DNA
duplex destabilization (SIDD) as functions of base sequence and stress level in long DNA
molecules. The base pair denaturation energies are assigned individually, so the influences of
near neighbors, methylated bases, adducts, or lesions can be included. Sample calculations
indicate that copolymeric energetics give results that are close to those derived when full
near-neighbor energetics are used; small but potentially informative differences occur only
in the calculated SIDD properties of moderately destabilized regions. The method presented
here for analyzing long sequences calculates the destabilization properties within windows of
fixed length N, with successive windows displaced by an offset distance d
o. The final values of the relevant destabilization parameters for each base pair are calculated as weighted
averages of the values computed for each window in which that base pair appears. This
approach implicitly assumes that the strength of the direct coupling between remote base
pairs that is induced by the imposed stress attenuates with their separation distance. This
strategy enables calculations of the destabilization properties of DNA sequences of any
length, up to and including complete chromosomes. We illustrate its utility by calculating
the destabilization properties of the entire E. coli genomic DNA sequence. A preliminary
analysis of the results shows that promoters are associated with SIDD regions in a highly
statistically significant manner, suggesting that SIDD attributes may prove useful in the
computational prediction of promoter locations in prokaryotes.
