COPICAT: COmprehensive Predictor of Interactions between Chemical compounds
And Target proteins
COPICAT is a system for predicting interactions between chemical compounds and
proteins by using Support Vector Machine (SVM), one of the most widely used statistical
learning methods. COPICAT realizes comprehensive prediction of protein-chemical
interactions by utilizing very general, or the most easily available, data i.e. amino acid
sequences and chemical structures.
COPICAT provides two functions; (1) prediction and (2) training. The 'prediction' of
COPICAT is made using SVM models. COPICAT default prediction models are based on
FDA approved drugs and their target proteins, including enzymes and GPCRs etc.,
according to the DrugBank database. In the 'training', users can upload a set of pairs of
a protein and a chemical compound with their sequence or structure data to construct
specific prediction models.
In the early-stages of drug discovery processes, the prediction of interactions between a
chemical compound and a specific protein can be of great benefit. Although 'docking
analysis', a 3D structure-based method, has been well studied and used in this field, it
suffers from the limitation of valid 3D-structure data and is time-consuming when
extensive applications, such as genome wide prediction, are intended. We hope that
comprehensively applicable COPICAT will contribute to the primary selection of
candidate chemical compounds from the vast chemical space and further to the
development of novel drugs.
|Sakakibara Lab., Dept. Biosciences and Informatics, Keio University|