Statistical Analysis Of Protein-DNA Interaction Site At Atomic Level

DNA-binding proteins play a pivotal role in various intra-cellular and extra-cellular activities, including transcriptional, translational regulation, DNA replication, DNA damage repair, etc. Protein-DNA interaction sites is an important branch in the field of bioinformatics.Generally, In order to predict protein-DNA interaction sites, we extracted the physical and chemical characteristics on the surface, such as amino acids, polarity, size and shape, etc. Then we created prediction algorithm based on these characteristics. Correct selection of the characteristics could improve the accuracy of predicted results.We compiled the dataset containing3D protein-DNA complex structure based on3D struction files of protein and DNA complex space in PDB database, also calculate each structure before and after all atomic interaction in the difference of the surface of the solvent accessibility. Thus we defined the interaction site of protein-DNA. According to the classification method of C, N, O, P, S, F, Cl, Br and I in Wang’s study, atoms is divided into22kinds of type. Frequency distribution proportion in different type of surface atom and interacted atom and the proportional distribution in atom type and average areas of atom type was calculated in our research.The results showed that22types of atoms appear in different probability in the interaction sites and no-interaction sites; The number and area of one atomic type on the interaction sites and no-interaction sites were significant differences. While the average area of every type of atom on the interaction sites were larger than that on the no-interaction sites surface.These neoteric results provide a possible way for predicting reseach based on protein-DNA interaction sites on atomic level.