Current Chemical Reactions In-House Database Is

Home , Melvin Calvin

CwYmnt@ommsntsQ EUGENE GARFIELD INSTITUTE FOR SCIENTIFIC INFORMATION* 3501 MARKET ST., PHILADELPHIA, PA 19104 International Collaboration Helps Establish 1S1 Reactions Chd): Current Chemical Rswctimr In-House Wtaba.re 1s Catalyst to Research Creativity

Number 13 March 30, 1987

Forty years ago as a premedical student Databases avaiiabie for personal-computer I encountered organic chemistry for the first USC.I For more comprehensive coverage, time. Like the thousand other students sit- we have developed an extensive chemical ting in that large Berkeley auditorium, I had reaction ffle for use on a mainframe com- learned simple inorganic reactions in the puter. Current Chemical Reactions In-House freshman chemistry course given by Profes- Database is a graphic and textual descrip- sor Joel Hildebrand. In that same lecture hall tion of over 5,000 new synthetic methods. I later learned about the wonders of Each record is accompanied by detailed bib- carbon-containing compounds from Profes- liographic and experimental data. The ad- sor Melvin Calvin, who would goon to win vantages of an in-house system, where the the 1961 Nobel Prize in chemistry for his software and databases are used on a com- work on photosynthesis. pany’s own mainframe computer, are dis- I retail from those days the endless hours cussed iater. spent memorizing dozens of organic chem- The student of elementary organic chemical reactions. At the time I wondered why istry may of course find it mind-boggiing there wasn’t an easier way to learn this new to imagine “memorizing” so many reac- subject, which I later came to realize was tions, but that is exactly what our new data- much like learning a new language. base does for you. We expect this file to Three years later I asked this same ques- grow to at least 50,000 new synthetic meth- tion regarding the actual preparation of ods over the next decade to facilitate the chemical compounds when I worked for identification of relevant means of synthe- Professor Louis P. Hammett at Columbia sizing new or known organic chemical University. Eventually I came to reaiize that entities. not only was there a better way but that a In writing about databases we tend to lose lot of others felt as 1 did. Information-re- sight of the creativity invoived in chemical trieval systems provide a soiution to the imi- synthesis. One might get the impression that tations of human memory. chemists are simply chefs executing cook- About 10 years iater in 1960, ISI@ book recipes found in a large electronic fde. iaunched the Index Chemkus@ (ZP ) to fa- But whiie artificial-intelligence programs cilitate iocating chemicai compounds in the have recently been developed in such sys- iiterahsrc. It was only a matter of time kfore tems as DENDRAL,2 LHASA,3 and we would tackle the more complex problem SECS,4 we are indeed far from replacing of the reactions that iead to these com- the human element in the creation of appro- pounds. In 1979 we began to index these re- priate synthetic methods for new com- actions on a systematic basis when Current pounds. Chemical Reacrionsa (CCR@) was born. After W. Tbeilheirner started the publica- Recentiy we discussed the hsdex f%vrricus tion of his ammal surveys of synthetic meth- and Current Chemical Reactions Personai ods in the mid- 1940s, the tieid of chemicat

83 information retrieval attracted a large num- Cuwent Chernid Reactions ber of creative people including R. Fug- mann, M.F. Lynch, W. Steidle, G. Vla- Our new iss-housedatabase is based on our dutz, W.T. Wipke, and H.J. Ziegler, to monthly publication Current Chemical Re- name a few. Another pioneer in this field actions. CCR provides a guide to new and is Jacques Vans, formerly with Rousell- newly modified reactions and syntheses re- Uclaf, Paris, but now at the Asian Institute ported in the current journal literature. Each of Technology, Bangkok, Thailand. In what entry contains a flowchart of the reaction as must now seem like an obvious truism, Vans well as complete bibliographic information wrote 13 years ago that “I am convinced and experimental notations when supplied chemists badly need a reaction doctrmenta- in the source article. tion [system] and when such a tool is put The synthetic methods reported by the at their disposal they [will] use it intensive- CCR are drawn from over 125 “core” or- ly.... Retrieving information on reactions is ganic chemistry and phmmaceutical journals of major—I would say of vital-importance from around the world. Four hundred new and at least as necessary as retrieving infor- synthetic methods are selected each month mation on chemical compounds.”s from this literature based on a strict set of criteria that screen for reactions or synthet- Reaction Data Club ic procedures that have not been published earlier; known reactions or synthetic pro- Many major chemical and pharmaceutical cedures that have been extended for use with companies worldwide have joined together different reacting groups; known procedures to support the CCR In-House Database. The that have been improved or modified; or the members of this Reaction Data Club (RDC) frostlaboratmy or total synthesis of a natural are listed in Table 1. They are a prestigious product. group of firms. Individurds from each firm will provide editoriai direction to the expan- The In-House Advantage sion and improvement of this database. The international cooperative effort by You may wonder why these files were not these major companies is not new, but this put up on a vendor system such as Dialog trend has certainly accelerated. The com- or Datastar as is &iSearch@, the ordine ver- plexities associated with the numerous sion of the Science Citation Indexn. We in- chemical products developed yearly prohibit vestigated possible vendors, but at the time individual companies from supplying all no vendor had the necessary programs to their own information needs. Therefore, for handle graphics-based reaction files in an the most complete and cost-effective way for online environment. However, the appro- companies to keep abreast of the chemical priate software for in-house systems was products developed worldwide, they and available. Making CCR accessible for in- their competitors rely on neutral third par- house mainframe use provides distinct ad- ties like 1S1to satisfy their needs. vantages. For example, access to the soft- It is for this purpose that 1S1took the ini- ware and databases is direct-it is simpler tiative to launch the Index Chemicus. Fol- and faster than dealing with a third-party lowing the same approach, we contacted vendor. In addition, it is possible to com- dozens of companies last year to find and bine data created in-house with 1S1data so guide this CCR database. Meetings were that a single combined search of both cart held in the US and Rome to detail our plans proceed simukarumusly. It is also possible for representatives from the 22 RDC mem- to create specialized databases. ber organizations. We expect membership It can be significantly less expensive for in the RDC will expand in the future, and a company with many employees requiring this will help control costs as well as enhance access to chemical data simply to buy the the database. sofiware and databases for installation in-

84 Table 1: Members of tbe Reaction Data Club.

EUROPE UNITED STATES Astra Laekernexlal Akdebolaget American Cymsamid Company Strangnaesvagan, Sweden Prketon, NJ

BASF Aktiengeaellschaft American Cyanamid Company Ludwigshafen Rhein Stamford, CT Federal Republic of Germany American Cyanamid Company Ciba-Geigy Limited Pearl River, NY Bad, Switzcrlmd Ciba-Geigy Corporation Hoechst Aktiengeaellschafi Summit, NJ Frankhwl am Main Federal Republic of Germany E.I. du Pent de Nemours & Company, Incorpxrtd F. Hoffmann-La Roche & Co. Wilmington, DE Baacl, Switzerland Hoffinam-La Roche, Incorporated Eli LiUy Nutley, NJ WindJesham, Surrey United Kingdem Eaatmasr Kwdak Company Rochester, NY E. Merck Darmstadt Eli Lilly & Company Federal Republic of Germany Indianapolis, IN

Pfizer Lhtritcd Monsanto Company Sandwich, Kent St. Louis, MO United Kingdom Polaroid Corporation RhOne-POulenc Cambridge, MA Saint-Fens, France The Upjohn Company Schering AG Kalamazoo, Ml Berlin, Federal Republic of Germany 3M (Minnesota Mining & Manufacturing Co.) St. Paul, MN house. Paying a third-party vendor for ac- CCR In-House &tabase. We chose the cess to the same information online is ex- REaction ACCess System (REACCS) of pensive if the file is to be used frequently Molecular Design Limited because we be- by many users. And there is always a con- lieve it is the system that best meets our cern about confidentiality. The questions one needs and those of our clients. Molecular asks on a public database ought to be pro- Design Limited is a h@ly reputable soft- tected by privacy laws, but recently certain ware company that enjoys about a 90 per- government agencies nave attempted to cent share of the chemical software market query online vendors about their customers. in the US and abroad. In addkion, REACCS The basic position of the information indus- can handle and display the types of reaction try, with which I concur, is that this would data in CCR and is used by more than 65 be a fundamental invasion of our rights companies around the world. under the Constitution. Certain European countries have already enacted strong pri- Features vacy laws. But apart from all this, those who are conscious of proprietary issues may pre- Beginning with data from 1986, the CCR fer in-house facilities. In-House Database covers 5,000 new syn- 1S1reviewed most of the candidate soft- thetic methods a year. However, since each ware packages available for use with the synthetic method has an average of five in-

$5 FSgure 1: Ssmple reaction record from the CC~ In-House Datobase,

o 0

THSCL 0 k iul HE(CH2)2CN

0 REFLUKED %- \ 681 YIELD ( wc) .hR-TUSHERONf

S.AKA1 T, MI VATA K. lSHIKAWA M, TAKEDA A. OK AYAMA u~lv, sc~ ENGN, DEPT 5yN CHEM, OK AYAMA 700, JAPAN, TETRAHEDRON LETT 26 {39),6727-30(19.95).

A NOVEL SYNTHETIC METHOD OF OIHYDRO–6-PYRDNE DERIVATIVES AND ITS APPL1C4TION TO THE SYNTHESIS OF AR- AT LA NT ONE AND 1+/- )- AR– TURMER0NE.

KEY RXN RING OPENING

REGNO : EXTREG: PATH : STEP: 4 30716 81 30F3

dividual steps, we provide a separate record In the growing information industry there for each individurd step of the new synthet- is a great deal of discussion about value- ic method as well as a step describing the -added information. The CCR In-House overall reaction. As a result, the database Database provides much more information provides access to over 30,0(N)reaction steps than that provided by the author in the each year. source article. ISI’s staff of professional Viewing an individual step out of context chemists enhance the appropriate data by from the entire synthetic method can lx con- highlighting both the reacting sites of the fusing. To ensure that you understand the compounds in a reaction as well as any ste- role of each step in the overall synthesis, the reochemistry involved if it is not clearly in- system provides a series of indicators. For dicated by the authors. On the computer example, the overall synthetic method may screen these areas are visibly enhanced by include the synthesis of more than one prod- color. If you don’t have a color graphics uct and result in various product pathways. monitor, the information is highlighted with The route for each individual step is shown bold-faced markers. Figure 1 shows a sam- on screen by a pathway indicator. The text ple reaction record with the three types of rdso provides a step indicator that tells you indicators and the reacting sites highlighted. the step number of the individurd reaction At a glance you can see that this sample as well as the total number of steps in the reaction is the third of three steps for path- overall synthesis. way B. Often a multistep method has only one step that is considered the new or criticrd Access ~OilltS part of the procedure. The text provides a key-reaction indicator to highlight this Each new synthetic method can be ac- pivotal step. cessed by chemical structure, and reacting

86 sites and stereochemistry can be specified. Suitable for a variety of industries in which In addition, each record has a large amount synthetic problems are encountered, this of supporting information that can be ac- database is a cost-effective aid for develop- cessed directly, ‘This textual material in- ing new chemicals and drugs while ensuring cludes bibliographic source data, reagents, that time-consuming duplication is avoided. catalysts, and solvents necessary for the In the computer world one hears a great deal reaction, If mentioned in the source article, about computer-aided design (CAD). There the conditions, yields, reaction descriptors, is a parallel to chemistry and scientific in- explosive reaction alerts, advantages, and formation. index Chernica.rused to be called limitations are also included. Current Abstracts of Chemistry & Index Chemicus (CAC/IC). CAC could, however, Conclusion convey a neat, alternative meaning-computer-aided chemistry, or indeed, comput- Since its introduction in 1979, Current er-aided creativity. Chems”calReactions has “beenmade available ** *** in print, on personal databases for use on My thanks to Lisa Holland and Daniel a Wrson~ computer, and now for main. Meyer for their help in the preparation of frame use with the CCR In-House Database. rhis essay. G,987 ,$1

REFERENCES

1. Gnrfield E. New compmmds and reactions are as close as your PC with the /rra’ex Chem”ca.r and Currenr Chemical Reactions Personal Databases. Current Contents (9):3-6, 2 March 1987. 2. BrrctrananB G, Felgenbaron E A & Lederherg J. On generality and prablem solving: a case stady using the DENDRAL program. Starrford, CA: Starrford University, August 1970. 50 p. NTIs/AD-715 128. 3. Corey E J, Hewett A P W, Jotrnson A P, Long A K, Mutzer J & Orf H W. Computer-assisted synthetic analysis. J. C)rem. In)orm. Corn#wt. Sci. 20:221-30, 1980. 4. Gund P, Grabowski E J J, Hoff D R, Sndth G M, ArrdoaeJ D, Rfmdea J B & Wipke W T. Computer-assistedsynthetic antdysis at Merck. J. Cherrr. Inform. Conrpw. Sci. 2088-93, 1980. 5. VaJfsJ. Reaction documentation. (Wipke W T, Heller S R, Fekimann R J & Hyde E, eds.) Corrrprrter representation arrd rrranipdation of ckmical inforrmztion, New York: Wiley, 1974. p. S3- 103.