bend.it® Server: Evaluation

How to evaluate a peak in the bendability/curvature propensity plot and how to describe the data.

The following phrase can be used in the methods section:

"The bendability/curvature propensity plot were calculated with the bend.it server, using the DNase  I based bendability parameters of Brukner et al. (1995) and the consensus bendability scale (Gabrielian and Pongor, 1996)."

Since the experimentally measured values of DNA curvature are quite variable, it is not advisable to use numeric values alone for the predicted curvature properties either. Instead, use a comparative qualitative description. This is more correct and also more understandable for the biologist. For example:

"The curvature-propensity plot, calculated with DNase  I-based trinucleotide parameters, contains one peculiar maximum in this region, whose magnitude (22.5) roughly corresponds to the value calculated for a highly curved motif in the Leishmannia tarentolae minicircle sequence. So we believe that this region may adopt a curved conformation."

"The curvature-propensity plot, calculated with DNase  I-based parameters did not show any conspicuous peak, however, the same plot drawn with the consensus bendability scale showed a conspicuous peak. In fact, a series of GGGCC motifs, similar to those found responsible for DNA curvature in vitro (Brukner et al, Nucl. Acids Res., 21, 1025-1029), was found in this region".

"The curvature-propensity plot, calculated with DNase I-based parameters, failed to show any value indicative of strong curvature (i.e. the values were all below 0.5; quite close to the genomic average value of this parameter). The bendability plot, however, showed a long valley indicative of rigid segments."


Andrei Gabrielian, Kristian Vlahovicek & Sándor Pongor