Advanced tree-based kernels for Protein Classification
Elisa Cilia and Alessandro Moschitti
The 10th Congress of the Italian Association for Artificial Intelligence (AIIA 2007)
Roma, Italy, September 10-13, 2007
Abstract
In this paper, we design novel models based on Support Vector Machines and Kernel Methods for the automatic protein active site classification. In particular, we devise innovative attribute-value and tree substructure computational representations derived from biological and spatial information of proteins. We experimented such models with the Protein Data Bank adequately pre-processed to make explicit the active site information. Our results show that structural kernels used in combination with polynomial kernels can be effectively applied to discriminate an active site from other regions of a protein. Such finding is very important since it firstly shows the successful identification of catalytic sites of a very large family of catalytic proteins belonging to a broad classes of enzymes.