No abstract

Abstract. We describe the Sumitomo Chemical Company computer program, which designs synthetic routes for organic molecules, without consulting a user. There are only approximately 2500 reactions stored in the program. Significant advances in the automation of reaction input are presented. We also give some examples of synthetic routes proposed by the program, one being an enantiospecific and stereospecific route for the synthesis of the antibiotic chloramphenicol. Brief historyThe father of the field of synthesis design using a computer was Vladuts who, in 1963, outlined plans and algorithms for the generation of synthetic routes, retrosynthetically, by a computer program'. Subsequently, Lederberg's group at Stanford programmed organic chemical reactions for the first time; this was in connection with the mass spectrum analyzing program, DENDRAL'. In 1969, Corey and Wipke submitted their first account of a pioneer program that could generate reactants via the user entering a particular product'. On iterating the process, an entire synthesis plan could be constructed as a product of a partnership between the chemist/user and the program. The chemist was required to express preference for one reactant or set of reactants over others, since it is impossible to consider all synthetic possibilities. In 1970, an account of the first program that planned synthesis by itself was submitted and this did not consult the user while running4. Our program5,", which we began in 1984, is a descendant of the latter program. We are able to run our program interactively, with a user prompting the program as to whether to proceed down the current pathway. However, in our hands, the interactive method has been less productive and, consequently, we don't use it. The style and environment of the interactive and non-interactive methods differ so greatly that we think it is clearer to use a different term for the non-interactive way, i.e., computer-executed synthesis planning. This contrasts with computer-assisted synthesis planning which presumes an interactive partnership. Gelernter's group' is also developing a computer-executed synthesis planning program. Model of synthesis planning process

The Sumitomo Chemical Company has begun a long-term project to build a program that proposes synthetic routes for organic chemicals. Preliminary results from the program indicate that the interactive method is feasible and preferable because of the superior speed of the computer in decision making. The routes proposed are generally quite reasonable; the efficacy of the program is limited by the fact that there are only about 2000 reactions in this early version.

Der Sumitomo‐Prozeß, für den Patentschutzbeantragt ist, liefert Sebacinsäure (IX) ausgehend von Naphthalin (I) anstelle des seither verwendeten Ricinusöls als Edukt, gemäß dem Formelschema.

No abstract