-------- Chemical Reaction Prediction
What is Reaction Prediction?
Our goal is to develop methodologies and systems for chemical reaction prediction in order to predict products and their ratios from given starting materials and conditions. This is the basis of the problems that synthetic chemists face in their daily experiments: will starting materials and reaction conditions give a compound possessing the desired structure, will the expected reaction actually occur, will the compound be produced as a major product, will a side reaction occur, what is the reaction mechanism, etc. Reaction prediction is one of the fundamental outstanding issues in chemistry.
Our Approach
Reaction prediction is quite difficult because of the complexity of chemical reactions that occur as a result of complicated interaction among many factors of varying degrees of contribution according to the reactions. Our approach is to establish a reaction map where chemical reaction data are organized based on physicochemical parameters concerning factors controlling chemical reactions. Using this map we wish to construct models that enable quantitative reaction prediction.
Classification and Prediction
The first and key step is classification of reactions and compounds. A first attempt was classification of 131 reactions based on changes of simple physicochemical parameters of reaction sites using a Kohonen neural network, which gave classification that had good correlation with the known reaction types.
For advanced reaction prediction considering more detailed factors, we developed a FRAU system, which characterizes compounds based on electrostatic and steric interactions with a pseudo-reactant. We classified reagent compounds based on the FRAU’s parameters and constructed a model that enabled prediction of possibilities of roles of the reagents, namely, structure transformation induced by the reagents. The results were verified with actual synthetic experiments.
References
- Satoh, H., Sacher, O., Nakata, T., Chen, L., Gasteiger, J., Funatsu, K. J. Chem. Inf. Comput. Sci., 38, 210-219 (1998)
- Satoh, H., Itono, S., Funatsu, K., Takano, K., Nakata, T., J. Chem. Inf. Comput. Sci., 39, 671-678 (1999)
- Satoh, H., Funatsu, K., Takano, K, Nakata, T., Bull. Chem. Soc. Jpn., 73, 1955-1965 (2000) (Headline Article)