BIOGRAPHICAL OUTLINE WILL TODD WIPKE PERSONAL INFORMATION Birthplace: St. Charles, Missouri Date of Birth: December 16, 1940 Nationality: U.S. Citizen Marital Status: Married, 2 children Military Status: Captain, U.S. Army, Honorable Discharge, September 1967 EDUCATION . B.S.: University of Missouri, 1962, Honors in Chemistry Ph.D.: University of California (Berkeley, 1965, W.G. Dauben) Thesis title: The Photochemistry of Transoid Dienes Postdoctoral: Harvard University, 1957-1969, E.J. Corey EMPLOYMENT Assistant Professor of Chemistry, Princeton University, Princeton, N.J. 1969-present. Postdoctoral research fellow in chemistry, Harvard University, Cambridge, Mass., 1967-1969. Automatic data processing and analysis officer, U.S. Army Combat Developments Command Air Defense Agency, Fort Bliss, Texas, 1966-1967. Research Chemist, Esso Research and Engineering Company, Baton Rouge, La., summer 1962, SCHOLARSHIPS Eastman Kodak Award, 1964 (awarded for research) NIH Fellowship, 1964-1965 Teaching Assistantship, University of California (Berkeley) 1962-1963 Texaco Outstanding Senior, 1962 (for research) Merck Career Development Grant, 1970 HONORS Sigma Xi, 1964 Phi Beta Kappa, 1962 Commendation Medal, U.S. Army, 1967 Distinguished Military Graduate, 1962 Omicron Delta Kappa (leadership and scholarship) 1961 Pi Mu Epsilon (mathematics), 1960 Sigma Rho Sigma (scholarship), 1959 SOCIETIES The Chemical Society, 1965 American Chemical Society, 1962 (Associate since 1959) Association for Computing Machinery RESEARCH INTERESTS Organic chemistry, structure and synthesis, photochemistry and excited states, chemistry of natural products, theoretical organic chemistry, ‘ organometallic chemistry and metal complexes, computer applications to creative chemical thought processes, analysis and chemical linguistics. W. Todd Wipke POSITIONS a Member, Member, Member, Member, Editor, Advisory Board of Chemical Abstracts Service (1970-1973) Editorial Board, Chemical Substructure Index, ISI Editorial Board, Computers in Chemistry, Pergamon Press Editorial Board, Journal of Chemical Documentation ial mation 4 Compal Season Computer Representation and Manipulation of Chemical Information John Wiley and Sons, 1973 Director, NATO Advanced Study Institute on "Computer Representation and Manipulation of Chemical Information," Noordwi jkerhout, Netherlands, June 4-15, 1973 Secretary-Treasurer, American Chemical Society, Princeton Section Member, American Chemical Society Committee on "The Role of Computers in Chemical Education." Member, Division of Chemical Literature Program Committee for American Chemical Society Member, National Academy of Sciences Committee to Determine Policy for the National Computing Laboratory > PUBLICATIONS 1. 2. 10. ll. 12. 13. 1h. "Photochemistry of Dienes," W.G. Dauben and W.T. Wipke, Pure and Appl. Chem., 9, 529 (1964). "Rapid Method for the Determination of Molecular Weight Distribution in High Polymers," L.W. Gamble, W.T. Wipke, and T. Lane, J. Appl. Polymer Sci., 9, 1503 (1965). "Nuclear Magnetic Resonance Spectra of Bicyclo (n.1.0) Alkane Deriva- tives," W.G. Dauben and W.T. Wipke, J. Org. Chem., 32, 2976 (1967). "“Computer-Assisted Design of Complex Molecular Syntheses," E.J. Corey and W.T. Wipke, Science, 166, 178 (1969). "Computer-Assisted Synthetic Analysis: Facile Man-Machine Communica- tion of Chemical Structure by Interactive Computer Graphics," E.J. Corey, W.T. Wipke, R.D. Cramer, and W.Jé Mowe, J. Amer. Chem. Soc., 94, 421 (1972). "Techniques for Perception by a Computer of Synthetically Significant Structural Features in Complex Molecules," E.J. Corey, W.T. Wipke, R.D. Cramer, and W.J. Howe, J. Amer. Chem. Soc., 94, 441 (1972). "Graphic Digitizing in 3-D," W.T. Wipke and A. Whetstone, Computer Graphics, 5, 10 (1971) "Computer Assisted Graph Theoretical Analysis of Complex Mechanistic Problems in Polycyclic Hydrocarbons. The Mechanism of Diamantane Forma- tion from Tetrahydro-Binor-S," T. Gund, P.v.R. Schleyer, P.H. Gund and W.T. Wipke, Computers in Chemical Research and Education, Elsevier, Amsterdam, vol. II (1973) pp 5/21-32. "Computer Searching of a Molecular Structure File for Pharmacophoric Patterns," P. Gund, W.T. Wipke and R. Langridge, Computers in Chemical Research and Education, Elsevier, Amsterdam, vol. II (197%) pp 3/37-38. "Congestion: A Conformation Dependent Measure of Steric Environmenc Derivation and Application in Stereoselective Addition to Unsaturated Carbon," W.T. Wipke and P. Gund, J. Amer. Chem. Soc., 96, 299 (1974). "Computer-Assisted Three-Dimensional Synthetic Analysis," W.T. Wipke, in Computer Representation and Manipulation of Chemical Information, ed. W.T. Wipke, S.R. Heller, R.J. Feldmann, and E. Hyde, J. Wiley, N.Y., (1974) pp 147-174. "Trans Addition of the Elements of Pd-Cl to a Diene," W.T. Wipke and G.L. Goeke, J. Amer. Chem. Soc., 96, hak (1974). "simulation and Evaluation of Chemical Synthesis. Computer Representa- - tion of Stereochemistry," W.T. Wipke and T.M. Dyott, J. Amer. Chem. Soc., 96, 4825 (1974). "Stereochemically Unique Naming Algorithm," W.T. Wipke and T.M. Dyott, J. Amer. Chem. Soc., 96, 4534 (1974). \ PUBLICATIONS CONTINUED 15. "Use of Ring Assemblies in a Ring Perception Algorithm," W.T. Wipke and T.M. Dyott, Computers and Chemistry, 1, 000 (1974 ) 16. "Computer Assisted Graph Theoretical Analysis of Complex Mcchanistic Problems in Polycyclic Hydrocarbons. The Mechanism of Diamantane Forma- tion from Various Pentacyclotetradecanes," T.M. Gund, P.v.R. Schleyer, P.H. Gund and W.T. Wipke, J. Amer. Chem. Soc., 96, 0000 (1974). Papers Submitted "The Oxidation of -Vinylcyclohexene with PdCl5/H20," W.T. Wipke and G.L. Goeke, Chem. Commun.. "The PdCls Complex of 4-Vinylcyclohexene," W.T. Wipke and G.L. Goeke, J. Organo- metallic Chem.. "simulation and Evaluation of Chemical Synthesis. Congestion: A Conformation Dependent Function of Steric Environment at a Reaction Center. Application with Torsional Terms to Stereoselectivity of Nucleophilic Additions to Ketones," W.T. Wipke and P. Gund, J. Amer. Chem. Soc.. Papers in Preparation "Computer Construction of Stereochemically Correct, Three-Dimensional Molec- ular Models from Two-Dimensional Structural Diagrams," W.T. Wipke, P. Gund, T.M. Dyott, and J.M. Verbalis. "Correlation of Congestion with Stereoselectivity of Olefin Epoxidation," W.T. Wipke, G. Birkhead, and T. Brownscombe. "Molecular Cartography. Application to Stereoselectivity of Radical Reactions," W.T. Wipke and A. Zelicoff. "ALCHEM: A Language for Representing Chemical Transforms. Application to Heterocyclic Chemistry," W.T. Wipke, C. Still, G. Grethe, T.M. Dyott, P.E. Friedland. INVITED LECTURES Symposium on Simulation and Modeling at IBM Watson Research Labs, N.Y. Conference on Uses of Modern Methods of Handling Chemical Information in Educetion of Chemists, National Research Council, National Academy of Sciences. Chemical Abstracts Services Sandoz Pharmaceutical Company Merck, Sharp, and Dohme Applied Math Program Western New York Organic Chemists Club State University of New York at Buffalo Hoffman LaRoche \ Union College General Electric Northwestern University National Science Foundation Rohm and Haas Up john . Rutgers University University of Tennessee, Knoxville University of Alabama, Tuscaloosa University of Georgia, Athens Tennessee Eastman, Kingsport Memphis Section of ACS UAIDE Symposium on Computer Animation in Chemistry, Miami Beach, Fla., October 12, 1970 American Chemical Society Lecture Tour, Jan. 4-9, 1971 Rudolph Anderson Symposium on Innovations in the Methods and Tools of Synthetic Organic Chemistry at Yale University, Jan. 14-15, 1971 Princeton ACS Section Columbia University University of Minnesota Givaudan Chemical Company Oberlin College Michigan State University Philadelphia Organic Chemists Club Computer Science Program at Princeton Sterling Winthrop Ohio State University Eli Lilly New York University Florida State University Chemical Society of Washington ARCO Chemical Company Pennsylvania State University Philadeiphia Section ACS, Toms River Subsection New York Section ACS, Westchester Subsection Ciba-Geigy University of Illinois, ACS Conference on Synthetic Methods, Carleton University, Ottawa, Canada Invited Panelist, "Computer Based Molecular Modeling Systems", Symposium on Computer Graphics in Medicine, Point Park College, Pittsburgh, Pa., February 22, 19/2. Invited Speaker “Heuristic Molecular Modeling", Symposium on Numerical Methods in Chemistry, National Institutes of Health, Bethesda, Md., March 20-22, 1972. (with P. Gund, J. G. Verbalis, and T. M. Dyott) Computer-Assisted Three-Dimensional Synthetic Analysis; presented at the 162 ACS National Meeting, Washington, D. C., September 12, 1971. (with T. M. Dyott) Fast Algoritm for the Perception of Rings in Chemical Structures; presented at the 7th Middle Atlantic Regional Meeting of American Chemical Society, Philadelphia, February 14-17, 1972. (with P. Gund, T. M. Dyott, and J. Verbalis) A Generalized Three- Dimensional Molecular Model-Building Program; presented at the 7th Middle Atlantic Regional Meeting of the American Chemical Society, Phila- delphia, February 14-17, 1972. Student Initiated Seminar Courses in Artificial Intelligence and in Computer Graphics, Princeton. Gordon Conference on Natural Products 24-28 July 1972. (with T. M. Dyott, P. Gund, and C. Still) Stereochemical Considera- tion in Computer Assisted Design of Organic Synthesis, presented at the .Stereochemical Symposium at the ACS National Meeting, New York, August 31, 1972. Washington University Medical School Monsanto University of Illinois at Chicago Circle State University of New York at Syracuse Rutgers Department of Computer Science McMaster University University of Toronto E. Merck, Darmstadt, Germany DuPont Up john University of Geneva Delivered lectures on Computer Graphics in Computer Center Lecture Series 4, Octoher 11, 1972, 1973, September 26 and October 2, 197k. Junior Colloquim on Applications of Computers to Chemistry, October 2, 1972 and October 9, 1973. Lecture to Computer Center Advisory Council, January 4, 1973. Invited Participant at "Frontiers of Organic Chemistry Symposium, " Hoffman LaRoche, Nutley, New Jersey, February 2, 1973. Stanford University Symposium on "Synthesis: A Science for All Seasons," November 12-14, 1973. Squibb University of California, Berkeley IBM, San Jose University of California, San Diego Shering Corporation Dow Chemical Corporation , Symposium on Strategies in Organic Synthesis, sponsored by Societe Chimi que de Belgique at the University Louvain la Neuvre Northern New Jersey Section of the American Chemical Society American Chemical Society Symposium on Pattern Recognition, National Meeting in Los Angeles Gordon Conference on Hydrocarbon Chemistry Gordon Conference on Scientific Information Problems in Research. “a -—% W. TODD WIPKE 1G KAw€ (975 __SUMEX-AUM QUISY OWE A EP Questions Relating to Medical and Computer Sciences Coals 1. Propesed research - see proposal, pane >. e. Present Support of this research: NIU RROOU(G Biowoloeiier Graphies Resource at Princeton, approved in November 97h for theo 0%) years q (sce Appendix I). Dr. Wipke is moving from Princeton to the University of California at Santa Cruz in August 1975. Because of this tove a new application for computing support has not yet been submitted. An experimental synthesis project for carrying out a SECS derived synthesis which occasionally requires computer conformaLlionil analysis is supported by NIN GM 20510-01. (ayjera'- Ta) Dr. Graham Smith, a postdoctoral fellow, is supported by an IBM /2llow- ship (salary $9000). | ep oay Dr. Hartmut Braun, a visiting fellow, is supported by Deutsche Forschungsgemeinschaft./ Ye 2% To! ~ + 3. The relevance of this research to the ALM of SUMEX is clear and strong. Organic synthesis is the source of many new medicines and desipnine a synthesis is a very difficult high level intellectual task requiring heuristic search and other techniques of AI. Our approach to this problem requires an interactive system with large meiioxry, large file space, high level languages, and support of remote interactive graphics. UCSC has only a small 3/0/1h5, and a PDP 11/5 (limit of 8 K words per user) available, both of which are unsuitable for this research. Collaborative Community Building 1. SECS will be capable of deriving syntheses of nearly any reasonable size organic molecule (less than 72 non-hydrogen atom). Additienally, we have demonstrated its potential utility in other areas such as mass spectril frag- mentations, molecular rearrangements, biogenesis, and biomolecular modeling. The SEMA algorithm in SECS which generates a stereochemically unique represent- ation for any organic molecular structure is useful to any biochemical inform- ation system. (See Appendix II). The stereochemical algorithms developed in SECS are going to be useful for enabling DENDRAL to generate stereoisomers. {lil *I o Applications progcams which I would usc in my work are the following: Program Name Use Source _ CONCEN For struct :ral elucidation Sumy ORTEP For structure plotting Cc... on* RE COG For pattern recognition LUI. PNMR For nmr simulation Wiy: WATKINS Nalf tone pictures Uta: DISMOL Molecular display and interaction Wipke STERIC Molecular cartography Wipke have DEC SYSTEM-10 version. 4. My programs have been made available after publication. for example, GIGL (General Interactive Graphics Language) is used ai. Princeton, Lawrence Livermore Laboratory, San Diego, Chicago Circle, and was also sent to Case Western. SYMIN (the synthesis model builder) was sent to Feldinann (NIN), Kowalski (University of Washington ), Jurs (Penn State), Chicago Circle, the DENDRAL projects, to PROPHET, and to Merck, Sharp, and Dohme. SECS has been sent to Merck, Sharp, and Dohme, University of Strasbourg, E. Merck, (Dermstadt ), BASF, (Ludwigshafen), and SANDOZ, (Basel). Additionally SECS has been placed on the First Data Corporation PDP-10 time sharing system in Waltham for access by everyone via teletype or GT 40 graphics terminal over dial up lines. This sharing of programs was possible because our programs are written in FORTRAN which enabled transfer to other DEC system 10s and other systems. The only difficulties I see in distribution of programs from our project viz “UMEX are: a) Tenex is not compatible with DEC system 10's. b) The large core on TENEX without traditional overlay decreases case of transfer to "normal" size machines. c) The usual difficulties encountered with large programs in getting them to run elsewhere. d) The need for an operator to mount, dismount and mail out tapes for remote SUMEX users. Na Yes Toam interested jn di seussing with SUMEX stafF interrelations of other AT work with mine. The AtM Workshop was a pood beginning of this, but T welcome additional suggestions Crom buchanan or Feigenbaum on specific problems we face in strategy and plan representation, There is very strong collaboration possible between DENDRAL and SECS - there has already been quite a bit of communication and sharing. 5. Yes, I already have a number of collatorations in synthesis and drug design and anticipate the interest in computer-assisted design of synthesis to stimulate more. | Hardware and Software Pequircments 1. I am currently using a PDP-10 (6) K) running 4s72 monitor (16 K) with an LDS-1 display which provides ):8 K memory For a user. We have: 1 GThO display 2 Dec tapes 1 Mag tape 9 trk 1 disk RPO2 teletypes Movie camera Cal comp plotter 3-D tablet met RN 2. FORTRAN is our current language. LISP or SAIL may be used on an exploratory basis, but because of transferability, we will probably continue to use FORTRAN... 3. I estimate my research group (est. 5 people) will require: a) cpu cycles: 5-15 hrs/weck. b) connect time: 3 £ull time dedicated lines at 1200-2100 band; each member of group currently uses about 20 hours/week connect time. c) at least one GTO terminal, a printer/plotter, and a couple other terminals. d) disk space of ~ 8000 pages based on the fact that one version of SECS requires about 1200 pages (source, rel., and binary) e) Offline media - will require program listings of about 100 pages /week; will require shipping about 2 DEC tapes per week (tapes to be returned). My group would need some but not all of this time during prime shift (0900-1600 PST). We would be willing to assist the community in the solution to loading problems, -- . Commniicab ion plems. Uesc is oF atv mites From SuHMX and is about 70 miles from San Jesse, TYHNEL commmnicatton via San Jose would be possible or leased lines directly to SUMeEX could be used, The basic cost of the latter is $150/month plus the cost of nodems. (200 band modem -$65/month; 1200 band modem $19/month). Vhere is also a one time installation charge of $100 and 462 respectively. IT would ask SUMEX to cover the communication costs until I can obtain a resource related prant. TI might need some support of a terminal de- pending upon what interim support I obtain. 5. This is a research project, but programs developed are expected to be of value to the medical comunity. I have already faced this problem with the pharmaceutical industry which wished to use SECS for design of drug syntheses - the solution was to place SECS on a commercial timesharing system so production work was removed from a research machine. PROVOSAT, ‘TO SUMEX-AIM RESOURCE Simulation and Evaluation of Chemical Syntheses (SEGS ) W. Todd Wipke y Introduction The development of new drugs and the understanding of how drug structure is related to biological activity depends upon the chemist's ability to synthesize new molecular structures and to modify existing ones or naturally occurring ones. Developing a synthesis requires knowledge of chemical transformations, knowledge of how structure affects chemical reactivity, and a capability for strategic planning in a global sense in order to guide scarch toward the goal. SECS is an interactive program for computer assisted planning of organic syntheses. We have focused on making SECS very knowledgeable about chemical transforms and chemical principles, letting the chemist do the majority of search direction interactively. The chemical knowledge of SECS is embodied in a library of chemical transforms written in ALCHEM, an english like language for describing chemical reactions including how reactions are affected by their environment. In order to evaluate certain chemical questions SECS builds its own three-dimensional model of the molecule and uses this model then evaluates the question in terms of the model. SECS also builds a model of the electronic structure of the molecule in order to evaluate electronic properties. Thus at its current level (2.0) SECS can evaluate with reasonable success the validity of the chemical inferences that it makes in working backward from the synthetic target to the generated precursors. However, its strategy for deciding what inferences to attempt is currently limited to a one step look-ahead, the look-ahead being limited to certain classes of transfoxins. Consequently, certain very interesting solutions are not found, because the chemist may fail to explore a precursor for which he cannot sce immediate potential value. Proposal This proposal is concerned with exploring synthetic stratcgies and their implementation in SECS. These strategies fall into several classcs: 1) strat- egics of how to simplify the target structure in general terms which correspond to long range plans; 2) strategies of how to achieve a given plan; 3) strategies of how to avoid obstacles in a plan (subgoal generation); +) and strategies of how to combine man and machine intelligence into effective interactive search strategies. These plans can be based not only on the graph theoretic proper- tics of the target, but also on the stereochemistry, proximity and steric rela- tions, and even electronic properties derived from the models which SECS constructs. I believe that strategies should be separate and largely independent of the specific chemical transforms. Only in this way can one experiment with strategies without modifying the chemical transforms. Similarly in this way, one can add new reactions to the reaction library without changing or even understanding the strategy knowledge base. This is quite different from the embedding of subroutine calls in a giant procedure representing the reaction, a technique that only reproduces the fixed sequences programmed in. The separation of strategic knowledge from chemical reaction knowledge becomes crucial as the knowledge base expands to eventually encompass most of organic or biochemistry. I believe it is the only way to preserve the “integrity of the knowledge base when many users are independently adding know- ] ‘ge as is now the case. The benefit of this research to artificial intelligence may be in the type }) strategies mentioned above and the interesting model and plan represent- Sa a ational questions raised. For chemistry, this approach will be generating new peneral strategies for the synthesis of complex organic molecules. It will also allow faster, deeper searches of synthesis trees which will serve the pharmaceutical chemist in designing new drug syntheses. Finally, I believe the framework of SECS will be applicable to other molecular task domains such as biogenesis and metabolism.