3D protein structure similarity comparison using a shape distribution method

Protein structure comparison is one of the most important problems in computational biology and plays a key role in protein structure prediction. A new method based on the shape distribution of Calpha representation of the backbone is proposed for evaluating structural similarity between protein pairs. The method is completely independent of sequence information. By examining the 3D proteins structural data from Protein Data Bank (PDB) using the method, strong shape distribution similarity is detected between protein pairs with close structures and functions. It provides a helpful method to measure protein pairspsila similarity when sequence similarity is not enough.