Identification of single-stranded and double-stranded DNA binding proteins based on protein structure

Area of Science:

• Structural biology
• Bioinformatics
• Molecular biology

Background:

• Protein-DNA interactions are crucial for biological processes but their structural mechanisms remain unclear.
• DNA binding proteins are categorized into double-stranded DNA binding proteins (DSBs) and single-stranded DNA binding proteins (SSBs), with distinct biological functions.
• DSBs regulate gene expression, while SSBs are involved in DNA replication, recombination, and repair.

Purpose of the Study:

• To investigate structural differences between DSBs and SSBs, focusing on surface tunnels and OB-fold domains.
• To develop a computational model for distinguishing between DSBs and SSBs based on these structural features.
• To enhance understanding of DNA binding protein specificity and function.

Main Methods:

• Analysis of surface tunnel characteristics (length, curvature) in DSBs and SSBs.
• Comparison of protein structures with OB-fold templates using maximal alignment score (TM-score).
• Construction and validation of a support vector machine (SVM) classification model.

Main Results:

• Distinct ranges of tunnel lengths and curvatures were identified between DSBs and SSBs.
• OB-fold domain alignment scores proved to be a discriminative feature for classifying these proteins.
• The SVM model achieved prediction accuracies of 87% (HOLO-set), 83% (APO-set), and 83% (Mixed-set).

Conclusions:

• Structural features, including surface tunnels and OB-fold domain alignments, effectively differentiate DSBs and SSBs.
• The developed feature set is effective for describing DNA binding proteins.
• The classification method demonstrates satisfactory performance on both DNA-bound and DNA-free proteins.