Development of a pipeline for generating ligand - protein binding site prediction tools using machine learning
Category
- Completed
- DICP Tool Prototype Trial for Integrated Database Analysis
- Projects funded in FY 2013
Name and affiliation of Research Director
BANNO Masaki
Postgraduate Student, Department of biotechnology, The University of Tokyo
Outline of R&D
We have developed a pipeline for generating tools that predict which residue binds a user-selected ligand in a protein sequence. We calculated all pairwise interactomic distances of protein ligand binding of atomic level in PDB database, and constructed ligand binding sites database. By integrating this database and other current biological databases, this pipeline generates high accuracy prediction tools based on machine learning method. These tools are able to predict the binding sites of a protein whose structural information is unknown at low calculation cost. Hence, this tool can be applied to genome wide prediction.
Main database(s) subject to research and development
- UTProt Galaxy
- PLBSPResidue
- UTProt CKAN
- Ligand Binding Residue Predictor Generator in UTProt Galaxy
Period of research and development
September 2013 to January 2014
