Further Development of a Reaction Generator in the SOPHIA System for Organic Reaction Prediction. Knowledge-Guided Addition of Suitable Atoms and/or Atomic Groups to Product Skeleton

A reaction generator in SOPHIA ( for reaction by ) generates all possible product structures by reconnections of free bonds obtained by cutting all bonds of a reaction site which is automatically perceived in the input reactantt structure and additions of atoms and/or atomic groups (AAG) to the free bonds. The reaction generator has been extended to automatically recognize and add suitable AAG for suitable free bonds by utilizing knowledge base derived from a reaction database. Contents of the knowledge base are structural information of AAG and structural characteristics of sites and their environments to which the AAG are added during reactions in a reaction database. The reaction generator considers reaction conditions to recognize suitable AAG for suitable free bonds by utilizing reaction condition groups obtained by classification based on word combinations of reaction condition descriptions in a reaction database. SOPHIA also has been extended to employ the reaction condition groups for interpretation of reaction conditions entered by the user and to consider the reaction conditions in reaction prediction procedures. This paper describes the contents and structure of the knowledge base, the reaction condition classification, and implementation of the reaction prediction using these results.