Analysing protein-DNA recognition

In subsequent work Mandel-Gutfreund and Margalit [12] used 43 known protein-DNA complexes to analyse the frequency of all pairs of amino acid-base interactions and construct a knowledge-based force field such as those used in the protein folding domain. They developed an additive scoring function based on the frequency distribution of the different pair interactions compared to the theoretical likelihood of their occurrence. The computed scores 

These studies have reinforced the idea that DNA deformation is an important component of the recognition process. However, as already remarked above, the number of complexes presently available excludes any analysis beyond the dinucleotide step level. This probably remains the major hindrance to developing a predictive model of sequence dependent effects. 

The application of PB theory to protein-nucleic acid complexes has resulted in two distinct lines of study. These approaches are characterised, firstly, by the work of Record and Olmsted who believe that using detailed structural models with atomic partial charges is unnecessary since the long range effects produced by the anionic phosphate groups dominate. They argue that the use of simplified 

Several important conclusions were reached in this study. Firstly, there is a deficit in the number of associated counterions for oligomeric DNA (in its native or denatured states) with respect to a polymer. This is due to electrostatic 

The role of changes in ligand conformation have been studied by Padmanabhan et al. [63] for five model oligopeptides binding to DNA. These 

---