Analyzing the Surface Structure of the Binding Domain on DNA and RNA Binding Proteins

The study of nucleic acid-binding protein (NBP) has important significance for us to understand critical intracellular activities, such as the transmission of cellular genetic information, cell metabolism, substance transport, and signal transduction. DNA-binding proteins (DBPs) and RNA-binding proteins (RBPs) interact through their diverse binding domains and different types of nucleic acid molecules. In this paper, we used a novel method that combines the CX algorithm and the fractal surfaces algorithm. This method gets the molecular volume and solvent surface area in the local area of the NBPs binding domain residues. Then, based on the algorithm results, the requisite domain residues are divided into three types: peak, flat, and valley. At the same time, we analyzed the solvent accessibility and secondary structural characteristics of the DBPs and RBPs binding domains. Finally, we found that there was an important difference in the distribution of peak residues and valley residues in the two types of NBPs binding domains. Similarly, there were significant differences in the solvent accessibility and secondary structural distribution of the two types of NBPs binding domains. To verify the existence of differences, we constructed SVM classifier to make a distinction between DBPs and RBPs using a 10-fold cross-validation method. Lastly, the SVM classification model achieves AUC of 78%. In summary, we have proposed a new perspective for the study of NBPs binding domains. This method not only calculates the geometric characteristics of the molecule, but also analyzes the protein properties associated with the structure, which will assist in the study of NBPs binding domains.