PUBLICATIONS: D. H. SMITH Page 2 11. . 12. 13. 14. ' 15. 16. 17. 18. 19. 20. The Lunar Sample Preliminary Examination Team, "Preliminary Examination of Lunar Samples from Apollo 12," Science, 167, 1325 (1970). D. H. Smith, "Mass Spectrometry, " Chapter X in Guide to Modern Methods of Instrumental Analysis, T. M. GOUW, Ed., Wiley-Interscience, New York, 1972. D. H. Smith, R. W. Olsen, F. C. Walls and A. L. Burlingame, "Real-time Mass Spectrometry: LOGOS--A Generalized Mass Spectrometry Computer System for High and Low Resolution, GC/MS and Closed-Loop Applications," Anal. Chem., 43-, 1796 (1971). A. L. Burlingame, J. S . Hauser, B. R. Simonei t, D. H. Smith, K . Biemann, N. Mancuso, R. Murphy, D. A. Flory and M. A. Reynolds, "Preliminary Organic Analysis of the Apollo 12 Cores," Proceedings of the Apollo 12 Lunar Science Conference, E. Levinson, Ed., M.I.T. Press, Cambridge, Mass., 1971, p. 1891. D. H. Smith, "A Compound Classifier Based on Computer Analysis of Low Resolution Mass Spectral Data,`! Anal. Chem., 44, 536 (1972). - D. H. Smith and G. Eglinton, "Compound Classi fi cation by Computer Treatment of Low Resolution Mass Spectra-Application to Geochemical and Environmental Problems, " Nature, 235, 325 (1972). D. H. Smith, N. A. B. Gray, C. T. Pillinger, B. J. Kimble and G. Eglinton, "Complex Mixture Analysis - Geochemical and Environmental Applications of a Compound Classifier Based on Computer Analysis of Low Resolution Mass Spectra," Adv. in Org. Geochem., 1971, p. 249. D. H. Smith, B. G. Buchanan, R. S. Engelmore, A, M. Duffield, A. Yeo, E. A. Feigenbaum, J. Lederberg and C. Djerassi, "Applications of Artificial Intelligence for Chemical Inference, VIII. An Approach to the Computer Interpretation of the High Resolution Mass Spectra of Complex Molecules. Structure Elucidation of Estrogenic Steroids,"' J. Amer. Chem. sot., 94, 5962 (1972). -- D. H. Smith, A. M. Duffield and C. Djerassi, "Mass Spectrometry in Structural and Stereochemical Problems, CCXXII. Delineation of Competing Fragmentation Pathways of Complex Molecules from a Study of Metastable Ion Transitions of Deuterated Derivatives," Org. Mass. Spectrom., z, 367 (1973). P. Longevial le, D. H. Smith, H. M. Fales, R. J. Highet and A. L. Burlingame, "High Resolution Mass Spectrometry in Molecular Structure Studies, V. The Fragmentation of Amaryllis Alkaloids in the Crinine Set-i-es, I' Org. Mass Spectrom., 7, 401 (1973). -- PUBLICATIONS: D. H. SMITH 21. 22. 23. 24. B. R. Simoneit, D. H. Smith, G. Eglinton and A. L. Burlingame. "Applications of Real-time Mass Spectrometric Techniques to Environmental Organic Geochemistry, I I. San Francisco Bay Area Waters," Arch. Env. Contam and Tox., 1, 193 (1973). D. H. Smith, B. G. Buchanan, R. S. Engelmore, H. Adlercreutz and C. Djerassi, "Applications of Artificial Intelligence for Chemical inference, IX. Analysis of Mixtures Without Prior Separation as Illustrated for Estrogens," J. Amer. Chem. Sot., 95, 6078 (1973). - D. H. Smith, B. G. Buchanan, W. C. White, E. A. Feigenbaum, J. Lederberg and C. Djerassi, "Applications of Artificial Intelligence for Chemical Inference, X. INTSUM - A Data Interpretation and Summary Program Applied to the Collected Mass Spectra of Estrogenic Steroids," Tetrahedron, 29, 3117 (1973). - G. Loew, M. Chadwick and D. H. Smith, "Applications of Molecular Orbital Theory to the Interpretation of Mass Spectra. Prediction of Primary Fragmentation Sites' in Organic Molecules," Org. Mass Spectrom. f z, 1241 (1973). . a7 SECTlON II -PRIVILEGED COMMUNICATION BIOGRAPHICAL SKETCH (Give the following information for ell professionel personnel listed on pege 3, bqinning with the Primipel Invartiptor. Use continuation pegar end follow the seme generel former for each psrronl NAME ITITLE IBIRTHDATE ma, OW, rd Sridharan, Natesa S. Research Associate 10/2/46 PUCE OF BIRTH (City, Stete, Country) PRESENT NATIONALITY (II nonUS citizen, SEX indicete kind of Vise and expiration data) India . ', Madras, India 5/73-U.S. permanent residence (g Male 0 Famals EDUCATION (&gin with baccelaurwte training end include postdoctorell SCIENTIFIC INSTITUTION AND LOCATION DEGREE YEAR CONFERRED FIELD Indian Institute of Technology, Madras Bachelor of India Technology 1967 Electrical Engineering State University of New York, Stony Brook M.S. 1969 Computer Science Ph.D. 1971 Computer Science HONORS University Fellow - 1968-1971, SUNY Stony Brook; Graduate Assistant - 1967-1968, SUNY Stony Brook; Siemens' Award (awarded for top rank in Electrical Engineering) - 1967, ITT Madras; National Merit Scholarship - 1963-1967, ITT Madras - MAJOR RESEARCH INTEREST ROLE IN PROPOSED PROJECT Computer Application in Chemistry and MedicinL RESEARCH SUPPORT (See instructions) Research Associate ---- RESEARCH AND,QR PROFESSIONAL EXPERIENCE (Sterling with pfscent posidon,list end experience relevant to eree of p.ro@cl, L!%t efi or most representetivepublicetioru. Do not exceed 3 peges for eech individual.) 1971-present Research Associate, Heuristic Programming Project, Stanford University 1970-1971 Consultant, IAC Computer Corp., Long Island, N.Y. Sridharan, N.S., "An Application of Artificial Intelligence to Organic Chemical Synti-'esis" Doctoral Thesis, State University of New York at StonyBrook, 1971. Sridharan, N.S., "Search Strategies of Organic Chemical Synthesis", Third International Joint Conference on Artificial Intelligence (3IJCAI), Stanford, 1973 Sridharan, N.S. (co-author), "Heuristic DENDRAL: Analysis of Molecular Structure", Proc. NATO Advanced Study Institute, Amsterdam, 1973. Sridharan, N.S. (co-author), "Heuristic Theory Formation", Machine Intelligence, Volume 7, Edinburgh, 1972. SECTION II - PRlVlLEGED COMMUNICATION BIOGRAPHICAL SKETCH NAME (Give the following Inform&ion for sllprofessional pononnel listed on page 3, brginning w'th the Principal Invertlgator. Use continuetion pqm snd follow the snme genaruf format for eech person) TITLE BIRTHDATE (Ma, Day, Yr.1 Brown, Harold D. Associate Professor July 12,1934 PLACE OF BIRTH Kiry, Sate, Country) PRESENT NATIONALITY f/f n0n.U.S citizen, SEX indicate kind of vis, and expiration date) South Bend, Indiana u.s, B Male Ofcrnale EDUCATION f8eoin wirh bacce!auraate ttaining and include oostdoctorsll lNSTlTUTlDN AND LOCATION University of Notre Dame, Notre Dame, Indiana Ohio State University, Columbus, Ohio (No Baccalaureate Degree) HONORS Summa Cum Laude - Notre Dame YEAR SCIENTIFIC CONFERRED FiELD 1963 Mathematics 1966 Mathamatics MAJOR RESEARCH !NTEREST %%k IN PROPOSEC) PROJECT Applied Discrete Mathematics - Computer Science Research Associate RESEARCH SUPPORT (See innrvctionrl Principal Investigator, NSF-GP-16793 (Expires March, 1974) Pending Proposal NSF (Proposed starting date September, 1974) RESEARCH AND/OR PROFESSIONAL EXPERIENCE (Starting withprarent position,&J&(&gandexpetien&rel@vant tort-e* i,fproht L.tiz or most representative publicetions. Do not exceed 3 pages for each individual.) Visiting Associate Professor, Computer Science, Stanford University , 1971-72, 1973-present Associate Professor, Mathematics, Ohio State University, 1966- Visitin.g Professor, Mathematics, Rhine,Westf. Tech. Hoch., Aachen, 1972 and 1973 Visiting Member, Courant Institute, New York University, 1967-65 Instructor/Assistant Professor, Assistant Chairman, Mathematics, Ohio State U.: 1963-65 Assistant to the Chairman, Mathematics, University of Notre Dame, 1960-63 Director or Associate Director, NSF-SSTP, 1964-70 Vitae Pace 2 7Ciea.r A`l@'L)ras, Ill. J. Kath, 12(1968), Pg. 215. Distrjbutor Theory in Rear AlZebras, Comm. Pure App. Eath. ??? ?o??O? ??? o ???? An Algorithm for the Determination of Space Groups, Ma& Comp. 23(w59), Pg. 499. Some Empirical Observations on Primitive Roots, with H. Zassenhaus, J, Number Theory 3(1971), Pg. 306. A Generalization of Farey Sequences, with K. Eahler, J, Number Theory 3(1971), pg.364, Basic Computations for tiders, Stanford CS Kemo STAN-CS-72-208. An Application of Zassenhaus' Unit Theorem, Acta Arith. xX(1972), Pg, 154. Integral Groups Tx The Reducible case, with J. Xeub:iser and II, Zassenhaus, Numer. Math. 19(i972), Pg. 3%. Integral Groups II: The Irreducible Case, with J, Neubtiser and H. Zassenhaus, Numer. Math. ZO(l972), Pg. 22. . Integral Groups IIIt Normalizers, with J. NeubGser and H, Zassenhaus, l-lath, Conp. 27(1973), PC;. 167. Constructive Graph Labeling Via Dollble Cosets, with L. Hjelmeland and L. Mnsinter, Di&rete Wath, in press and Stanford CS I:e?lo STAN-E-72-318. An Algorithm for the Construction of the Graphs of Organic liioiecljles, with I., Kasinter, Discrete Hatll. in press and Stariford CS E:eno STAN-G-73-261, ) The Crystallogaphic Groups of &-dimensional Space, with 2, Neubcser, H. Wondratschek and H. Zassenhaus, Wiley-Interscience in Fess. 30 SECTION II - PRIVILFtiEO COrulMUNICATlON --- ------ 8lOGRAPHICAL SKETCH -II {Give the following information forsllprofmsionel personnel listed on page 3, brglnning w'th the Princi@ Invartlpgtor. Use continuation moos end follow the s?rne aenerrtl former for eech person) NAME - TITLE BIRTHDATE (Ma, Day, Yr.) DROMEY, Robert Geoffrey PLACE OF BIRTH (Ciry, Stste, Countryj Research Associate 11121146 PRESENT NATIONALITY (If non4.S citizen, SEX indicete kind of visa end expiration de tel Castlemaine, Victoria, Australia ustralian, J-l Visa, Exp. 10/8/74 ~~~~~ o>cms'a EDUCATION (Begin with baccelaureate training and include postdoctoralj SCIENTIFIC INSTITUTION AND LOCATION DEGREE YEAR CONFERRED FIELD Swinburne College of Technology, Diploma of 1968 Chemistry Melbourne, Australia Appl. Chem. La Trobe University Ph.D. 1973 klolecular Science Melbourne, Australia HONORS CSIRO Postdoctoral Studentship Commonwealth Postgraduate Research Scholarship Walter Lindrum Memorial Scholarship Artificial Intelligence Techniques to Bio Medical and Chemical Problems. RESEARCH SUPPORT (Sseinn~ctionrl Research Associate ---- RESEARCH ANDjQR PROFESSIONAL EXPERIENCE (Starting withpracent positicn,~~sndexpen'sncerel#ent foor~s ofpmbt Lind ormort mpnasntetivepublication+ Do not exceed 3pages for sach individual.) 1973 DENDRAL Project, Stanford University, Computer Science Department 1973 Software Development for Graphics Systems, LaTrobe University, Computer Centre 1969-73 Construction, development and applications of an on-line photoelectron spectrometer LaTrobe University, Chemistry Department 1569-73 Application of Deconvolution Techniques to the Processing of Experimental Data. Publications: "Deconvolution and Its Application to the Processing of Experimental Data", Intl. Journal,of Mass Spectrometry and Ion Physics, 1970, 5. (co-author). "Inverse Convolution in Mass Spectrometry", Intl. Jnl. Mass Spec. Ion Phys.,1971, 6. (co- author). "A Combined Time Averaging-Deconvolution Technique Applied to Electron Impact Ionization Efficiency Curves", Internation Journal of Mass Spectrometry & Ion Physics, 1971, 5. (co-author). "The Perfect Direction and Velocity Focus at 254034' in a Cylindrical Electrostatic Field", Reviews of Scientific Instruments, 1373, 44. (co-author). - R. G. Dromey *'Detection of Spin-Orbit Splitting in the Photoelectron Spectrum of 02+ by Deconvolution", Chem. Physics Letters (in press), 1973. (co-author) "The Effect of Finite Line Widths on the Interpretation of Photoelectron Spectra", Journal of Electron Spectroscopic (accepted for publication). (co-author). "An On-line Ultraviolet Photoelectron Spectrometer for High-Resolution Studies o-f Molecular Structure", Australian Journal of Chemistry (accepted for publication). (co-author). "Photoelectron Spectroscopic Correlation of the Molecular Orbitals of the Alkanes and Alkyliodides", Journal of Molecular Structure (submitted for publication). (coauthor). "Comparison of the Photoelectron Spectra and the Photoionization EffFciency Curves for the Alkyliodides", Transactions of the Faraday Society (submitted for publication). (co-author). "A Convolution-Deconvolution Algorithm Using Fast Fourier Transforms), Decuscope, 1973 (in press). RESEARCH PIAN 33 I. rr, ITT, ?? o ?? ? ?? ? PTOMOLECTJLAR CHARACTPFT7AfI3N: AFTTFICTAL INTFLLIGENCE A Program of Fesource-!?elated 3esearch A. 3bject.ivcs n. Backqround artd Rat ionale C. Pelationship to AIY-SUMFX and the Genetics Research Center SPECIFIC AIFlS t3FTH3DS SIZNIFICANCE OF PROPOSED FESEARCE FnCTLITIES 8 Z;QUIPI?FENT ETBLIOGRAPHY SahL~ 1 Fiqurps l-3 Appendix 4: Letters of Interest 4op enlix R: 1973 Annual Report to the NTii 1. INTFDDUCTION Thi 3 r?czec3'L application is intended to susfain and augment the zapabilities of the mass spectrometry (MS) program which has ;ecv~!il 3s 3 major insTitutional resource at Stanford for some yp3rs. Rith previous support from NASA and NSF it has made oossible a hiqhly interdisciplinary set of research projects r;tnqiPq ov2r: artificial intelligence (AI) in biomolecular ~h3r3cterization, natural product chemistry, clinical biochemical sttl!ies on steroids, and the mechanisms of molecular fragment Formation in mass snectrometry. While the facility equipment for mass spectrometry has been funded mostly by other agencies, cDnne!cted research proqrams embrace several NIY research projects is wzll, In addition, this activity was closely coupled with the ?,"FF Ye!ical school computer resource (1966-19731 and will have sisflnr associations with the new RI?!-S!JEl?X conputer resource re:ontly flnded by the BP2 (see Section T."). Previous support ref1ect.s the diversified facets of this icperdisciplinacy research, RASA has supported projects in new i?:trunent3tion, inclurlinq the initial mass spectrometer-computer Ilnk, gSF has supported chemical research, and ARPA has supported our artificial intelliqencc research and initial application to nass spPctrometry, Overall cutbacks have forced NASA to reduce Ftinqinq for this area of research despite their interest. Under AFP.4 sunport to Drs. Feigecbaum an3 Lederherg for AI research, the DENDPAL proqram became recoqnized as one of the most successful AI 3pplic3tions programs. Eowevec * ARPA is chartered to fund Fr3nti~c computer science research and no longer provides funds fnr the DENDRAL applications programs. APPA has indicated a relozt3nz~ to continue funding to this groun for the theory fncmatior, work in chemistry, although we expect to continue to receive AfiPA support for more theoretical aspects of OUT research 3r3ffr451 (e*q*, 3utomati.c proqrarrming). iJc previously submitted a comprehensive proposal to the NIH (Rt?-~07SS, 3/25/73) which included an application for the 4fY -SUVEX zr>mputinq resource and a renewal o f the existing DENDRAL qrant IPP-00012) o This proposal was approved for 5 years by the Pational Advisory Researc+ Fes3urces Council. Certain reservations were, however, communicate3 to us: they concerned especially what we must aqree was an ambitious effort to close the control loop for l'intelliqent. automationfi* whose costs overreached t he imme~i3~e utility of the expected result. During subsequent 3iscussiDn? with the RioTechnDLoqy Resources Branch, takinq into accnurt the council review aR.d a number of diverse policy issues, we 37ceoit administratively to seqment the two components of the 3riqinaf proposal. The ATPI-SU?!EX portion of the original proposal (exclil3inq DI?F?DRAL) was recently funded for 5 years as a national resource for artificial intelligence in medicine. The present ~r~p~ri31 for resource-related research in hiomolecular characterization and artificial intelliqence is an elaboration of thz DEN?RAL portion incorporating intensive reexamination and revision of the previous proposal. Vith thr? differentiation of priorities represented by AItl-SUEEX, th? "tlnetizs Fesearch Center (ZPC), and continuinq work on Irtifizial intelliqecce under Dr. Feiqanbaum's leadership, the Dresclt renewal application places more emphasis than heretofore 3n real-world oriented applic2t.ions. Correspondingly, we have sqceei that it is now more appropriate thst Dr. Djerassi should be ?ps!qn3ted as Principal Investiqatz)r in this phase of our work. Fs outline3 in section 3.2, the interests 2nd responsibilities of ?r3f?ssors Djerassi (Chemis+ryl I Feiqenbaum (Computer Science) and '_e?erberq (Genetics) have? been closely intordigitated. With their f!lrther connections with many colleagues, these programs enjoy a h irJh acqrs? of university-wide participation. For example, the ?zRttics rtppartment is also closely affiliated with Biology, Bi 3zhemis+.ry, Pediatrics, Psychiatry an3 xedicine through joint lppointrnents or joint research projects or both. This breadth would be difficult to obtain except at a few institutions where t5s medical. schDo3. is both aca4emicslly an3. qeoqraphically inteqrlted with the university t:, the deqree that characterizes the Stanford University envir0nmeF.t. GLOFSARY !IF ABDREVIATIO?JS AC"F - Advanced Computer for Yedical Research {Nib-funded computer resource, 1968-1973) 91 - artificial intclliqencp 4 It!-S!Tr!??Y- A comprehensive computer resource jntended to serve the n&iqnal requirement for artificial intelligence in medicine. This will be implevsnted at the Stanford University facLlity called AIM-SVYEX - Advanced Research Projects 4qency of the Department of Defense. - Biot?chngloqy Resources Rranch - carhDn-l? saqr;etic resonance - qas chroaatoqraphy or qas chromatoqraph - Genetics Research Center (Stanford, J. Lederberg, Principal Investiqator; NIGYS-approved and awaitinq fundinq. Grant #PDl-GY 20832-01) - hiqh resolution mass spectrometry - infra-red - Tnstrumentation Research Laboratory {Stanford Gene%ics Department) - low resolution mass sppctrometry - magnetic circular dichroism - Mass spactronetry or mass sDectrometer - National Aeronautics G Space Administration - nuclear maqnetic resonance - National Science Foilndation - optinal rotatory dispersion - a mo-lified version of the PL-1 computer language (for the Stanford ACME computer fscility) - Stanford University Eedical Experimental Computer Resource (WIH funded computer resource, 1973-1978) - ultra-violet 3& A. 03JECTTVES: Core Research. "he fllnds now applied for would permit 1) the continued fundinq of the ?!S laboratory as a biomolecular oh3r37cerization resource; ?) advancement of laboratory instrumentation capability in specific areas of GC-HRMS and the exploitation of metastahle peak linalvsis. 3) the fucther development of AI computer techniques to match the instranentation. This work will emphasize practical utilization for applications in biomolecular characterization connected with other on-qoinq biomedical research programs. It will include, for 3X3UP1, al the analysis of mixtures by SC/MS; b) metastable peak 3n3lvsis f>r difficult problems of pure compounds and of mixtures not rcalily separable by CC; c) optimized data analysis for zhsr3cterization of MS peaks an3 d) heuristic analysis of spectra for the molecular ion composition. ?ur project is the only systematic effort, to our knowledge, cl~rrently underway in this country for computer assisted structure ?licidation, Subsequent to our early publications, an intensive proqram has been mounted in Japan in similar areas. This situa+.lon may ho contrasted with computer assisted organic synt.h,+sis, an area receiving considerable attention from several ccse3rzh qroups, The%? capabilities can be beneficially provided t3 a wider community via the ATM-SUP?SX resource. Research on the ?nulstion of human intellect by computer programs will undoubtedly Influence the efficiency with which chemical research can be soplied to ever more complex problems of health, e. q., intermediary metabolism and its pathologies; environmental infl3F?n:a s on hea1t.h; the development and critical validation of new therapeutic aqents. "he a:Elevament of these nbjectiQeS depends on the continued maintenance and development of the DENDP9L AI programming system {see h313WI . The advent of the ATM-StMEX facility will remove some nf the serious computational limits on the exercise of this system that have delayed recent progress. Pduc3tion. rn our university setting, pre-doctoral and post-doctoral educa+_ion of course constitutes a part of our mission. As far as is practically possible, research participation in the DENDRAL proqram has been couple3 with 3 issertation w3rk by graduate stu??nts and post-doctoral research experience respectively. Ex3 mplss of people (and their research area) whose education has Seen enttan:e d in this way are the followinq: Srad~late Students: J, Simek, padaqoqical aspects of the structure qerlerstor; Rai Lee Tan, synthesis of new estroqen compounds; H. Fqqcrt, 13ZYYIR of amines and st?roidal ketones; C. Van Antwerp, 13CYR of steroidal alcohols; C. Farrell, theory formation from ~a3s spectral data; L. !!asint.cr, development of the structure qnnnrator; Y. Stefik, AI applications to chemistry. 37 F'nst3933rll Fcllous: r,. Drr)mey, theory formation f ram analytical !3t3; P. Sritter, mass spectral fragmentation of bioloqically active steroids; R. Carhart, analysis of 13CI?R spectra by DSNDRAL-like programs; 5, Hammerurn, development of better fraqmcntation rules for proqesterones. Fofmll orqlnization. *his project has been a long-term commitment of Djerassi, Lederberq and Feiqenbaum functioning in effect as Jo-investigators, We coordinate our activities with day-to-day zDntlc?s i.1 the pursuit of converqent research objectives. In the light of the extension of our collaborative activity during the 1.3;t tdo years, we are now orqanizinq a formal advisory group to i ~cludp, in addition to ourselves, H. Cann, J. Barchas, and E. Van Tamelen, This qroup will advise the principal investigators on the direction of the proqram with respect to allocating available facilities and seekinq out and helping other collaborators, This desiqlation simply recoqnizes the fact that many of our colleagues haIre already been ecqaged in c. alevant collaborative research with 1s. A '?S resource has recently been funded at the riniversity of California/Berkeley, under the direction of Dr. A.I.. Burlingane. Drs. Djerassi and Burlingame have recently engaqed in some zollahorative research which was made more successful by the sharinq of facilities ard expertise available at or.e institution but n?t at the other. We would hope to maintain and strengthen t hES2 contacts to avoid unnecessary duplication of effort. U? plan to discuss with Dr. A.L. Burlinqamo the most appropriate pr3c?Aures for coordinaticq the related activities of our respect ivn proqrams at the 1Jni.versit.y of California/Berkeley and !I?_?. ?hi; may take the form of reciprocal membership in advisory committees. fhs '*hardware resource w to which this application is pegged has 333~ id entified as the MS facility. ahile these instruments alone represent an investment of over 5`300,000, funded previously by several aqencies, they do not represent the most important resource. WC, would use this designation instead for the working team led by th,e princioal and co-investiqators. The skills embraced by this group Fncllde, as mentioned, computer science, structural orqanic zhemis+rv, molecular biology, instrumentation enqincering and a wide ranqe of other disciplines, They are represented not only in tke princinal professors hut in a diversified and accomplished orofessional research staff (SE?;? Bu3qet Justification). The proqrlm for which funds are now requested is the vital means by which the interests of this qroup can he sustained in a zn?r3inate? effort that would be very costly both in funds and in tin4 if it had to he reconstructed from scratch. 'dithout the finanziatl support now requested, this line of collaborative research will have to he abandoned, with it a unique style of interdisciplinary collaboration, and the "1s facility will be terminated. 9. BACRGP3JJND AldD RATIOWALE 1, Phe StrtlctVlre Elucidation Problem 3) The General Problem, Analysis of molecular structure is a major activity in our program of resource related research. For the speci.fic task of efuci3ating molecular structures, i.e., the topoloqy of atom-to-atom connectivities, analysts utilize a mixture of information derived from chemical procedures and spectroscopic techniques. Each item of information, if not red!lndant 3r uninterpretable, contributes to t4e solution of the problem. Chemists 3raw upon a tremendous body 7f snecific knowledge about the task area (e.q., clinical chemistry, biochemistry), molecular structure, spectroscopic techniques, etc., in or3er to piece together this information and l?fer the structure of aolecules. These features, and the relative si!nplicity of the finai concept of a structure, make the prohl?m particularly well-suited for applications of the tec!iniqtles of AI to assis t research workers performing the task. bl D-jerassF*s Laboratory. Professor D jerassi has been concerned with secncture elucidation problems since the beginning of his -hamica research. Y kL,d His activities at Stanford have been concerned heavily with the apnlication of particular spectroscopic techniques to structural studios of biomedically important ~3nP3unas. These techniques include optical rotatory dispersion IoR?) and, more recently, maqnetic circular dichroism (BCD) (both 3f t.hem supported initially by the NIH), Since 1961 he and his ?coun have also been concerned with MS because of the power of the technique, in terms of specificity and sensitivity, as an an3lyticsl tool for structllre elucidation. Four books and anproximately 250 articles on 9s have been published by him and i i.s zolIea7ues. rha technique of E4S does not silffice for all structure 3zt~rminatio~ problems, but it is a very powerful tool in areas where thora exists a body of knowledge about the MS behavior of related molecules, Bhen sample size is limited MS may well be %he only technique that tag be utilized. The recent availability of hiqh resolution mass spectrometers has madn HFMS the technique of c: h 3 s z e four ntany applications because under ideal conditions the exact mass number uniquely specifies the the empirical formula of 3 molecule or fraqraent. nn a parallel course, the technique of ;:/?pS, routinely available with low resolution mass spectrometers (r,C,'LRMS), has revolutionized investigations wherever complex aixtclrss ace encountered. All of the above considerations argue that an extension o f H,S at Stanford to provide routine GC/LRMS and ;C/!fRMS an.alyses would he the next logical step to assist researcher; depending on this facility for solutions of their structure elucidation problems. 2. Aistorical Backqrnund 91 Y3ss Spectrometry Laboratory. Prior to the existing DENDHAL 7 r I c t , the qroundwork was laid for computerization of the existing mass spectrometers, an Associated Electrical Industries MS-9 high resolution mass spectrometer and an Atlas Z'i-4 low resolution mass spectrometer. This work, supported primarily by NASA via the Tnstru~cntation Research Laboratory (T9L) in the Department of :?r!etiZS, resulted in link-up to the then existinq ACME computer facility via a PDP-11 mini-comolIter which acted as a buffer bctweon tha spectrometers and ACHE. Initial data acquisition and ceduztion proqrams were written for thz system and utilized on a limitz?d basis, The fundinq of the DENDRAL proposal, NIH qrant PU-612 [May 7,1Q71-present) in conjunction with additional resources provided by the IRL resulted in a major improvement to these capabilities. The fruits of these? efforts are described un3et section T-B.3 [below). h) Summary of Early DEYDRAL Development. r? 1364, Lederberq devi.scd a notational algr,rithm for chemical :):r!lz tllr?s (termed DENDPAT,,f that allowed questions of molecular structure to b? framed in precise qraph-theoretic terms. (Refs. 1,3-5,12)* He also showed how to use the DENDPAL algorithm to qenerate complete and irredundant lists of structural isomers. (RCfS. ???? o Tn lQ65-66 Lederberq and Feiqc?nbaum began exploring the idea of usin th: isomer qenerator in an artificial intelligence program - searchinq the space of possible structures for plausible solutions to a problem much as a chess-playinq program searches the space of lesal moves for the best moves. (Refs, 7,12). This approach qusrsnt??s that every possible sollfion to a problem is considered - 2ith9r ianlicitly, as when whole classes of unstable structures are rz-jectsd, or explicitly, as when complnto molecules are tested for plansibility, In either case, an investigator easily ?etPrmines the criteria for rejection and acceptance and knows that no possibilities have been foryottpr., l!his approach also guarantees that structures appear in the list only once - that 3utor3rphi.z representations of the Sam? corrplex molecule have not h?en included. In both these respects the computer proqram has an advantaqe over manual approaches to structure elucidation. Cl Initial collaboration with Djerassi. (Refs. 14,15,19, 20,21,22,24). Ledetberq and Feiqonbaum realized that (a) only throuqh aaplicatio? to real problems could the AT approach be materially l?v,ancad and critically evaluated, and (b) MS appeared to be a fruitful applications area. MS appeared to be an excellent orohlem area because of the close relationship between spectral Tr3qmen+atiop patterns and molecular structure for many classes of a?leculls* Djerassigs interest and expertise - an3 daily intpr3ctio2 between members of his qroup and the AI group - led to 9 seriF5 of joint publications describing the approach and initial cesrllts of the programs. The success of these collaborative ?)lf>rts led to the proposa?. to the NIH for initial fundinq to ex+e?I these efforts. 1) Efforts Under NIH Funling for DENDRAL. (3efs. 25-41). ?h? initial fundinq by NIH provided the opportunity to upgrade the Fnstrument~tioc and computer pcoqr2ms. In particular we were able to !llnun+. a concerted project on both the analysis of mass spectra of bioTedically important compounds and the mathematical aspects of m31?cul3r structure. Proqress reports to the NIH describe this research in detail. The nest recent annual report appears in 4 r?pe!li?x 5. R series of publications Sirccted to audiences both in corsputer science and chemistry sre listed in the bibliography. Ilh? followFnq section [Section 3) summarizes the capabilities for -tcllcture elucidation which, 3 in thsmselves, constitute an impqrtnnt result of past wnrk. e) Related Research. Fr! important side effect of tha DEWDRAL project is the extent to vhich additional research was inspired and carried out to fill q3cIs in existinq knowledqe. This research, not supported by the r)Srt'DRhL qrant, has been beneficial to or-qoicq DENDRAL work, and vine-versa. Publications whjch have arisen from this research are listed in the hiblioqranhy (Pefs, 58-70). A brief review of these oublications should indicate the need for precise specification of the kn7wledqe elicited from chemists and used in computer Dr3nr3ms. As an example, consider the description and application >f an early alqorithm for qeneration of cyclic structural isomers (311. This paper considered the problem of spectroscopic diffar?ntLation of isomers of Z6H100. Unsaturated ethers fall in 3ne of the classes of isomeric compounds which must be considered, but th? RS of unsaturated ethers had not been investigated systematically. This work WAS subsequently carried out in Professor Djerassi's laboratory independently of DENDRAL support, but of hen?Eit to DZ':NDRAL (62). Other examples will be found in t h3 Bihlioqraphy (Eefs. 58-70). 3. Existinq Capabilities an hlvp worked to develop distinctive capabilities for molecular structure elucidation, bringing together a hiqh quality HRMS svstea and AI programs applied to biomolecular characterization. The feasibility of our analytical approach has been demonstrated in saveral problem areas, base3 upon the development both of a MS svst3v ;Init a qeneral se+ of computer proqrams for use in new areas. The princioal capabilities are summarized below. These are now in 3pilq and Here developed primarily under NIH funding to this pc3 jer=t, with additional support supplied by ARPA and NASA in snecific areas. (These agencies have reduced funding levels for this work because overall cutbacks have forced NASA to cut out this are3 3,f research despite their interest and ARPA is chartered to provide funds for frontier coapllter science research but not f9c 3pplic3tions. Thus the NLH is the principal of support for future development of anplications programs in the interdisciplinary area of artifici31 intelliqence./health related r:herrical problems.) 3, HRYS System and Coupled SC/LRMS System. We h3ve coupled the PIH-supported Varian -YRT 711 High Resolution Yass Spectrometer with a Hewlett Packard Gas Cbromatograph and denonstratad its utility for GC/LR?!S analysis of such difficult analytic problems as the free -;terols (i.e., not derivatized) isolated from marine and other sources. Advanced data reduction techniques for this instrument Yere written for use with the ACME conput er svstem (36Q/50) and cow exist in Stanford's new 370/158 ihI. Ch :Jnntinues to support the PL/ACNR lanquaqe. nJC/HRMS scans on extracts from urine and anniotiz flui demonstrated this system's zapability to provide hiqh quality mass measurements on complex nixtures obtained from biological sources. An example of one ;t/fiRYS run on the amino acid fraction of amniotic fluid is presented below (Sec. 1II.D). DEWDRAL Structure Generator (Refs. 1-6,14,31,37,38,40,41) nhe D;iDRAL Structure Generator program accomplishes exhaustive ; an? irredundant generation of isomers, with and without rings. This pcoqram guarantees consideration of every candidate structure - either implicitly, as when whole classes of structures are focbi3"en, or explicitly, as when individual compounds in a class -1ro spec<,fied, It corresponds to the "legal move generator" of c3nnutcriz2d chess playinq and other heuris+ic programs. C. DRNDRAL Planner (Refs. 25,28,33) de have written a very qeneral set of computer proqrams for determining structural features from analytical data in well-defined areas. Such qeneral planning programs have been wri+t?n fo: low and hiqh resolution MS, interpreted proton NJ4R swectroscopy and 13CMR data, 3, INTSUM (ReFs. 26,29,34,35) TNTS`JM ?s 3 computer proqram that aids in finding interpretive rales for YS. The proqraa interprets a larqcl collection of MS data 1:: orjinq to criteria specified by the investiqator. Then it summarizes the data to show which of the possible interpretations seen most plausible. P. PllLEGEhr (Refs. 26,35) RITLESEW is the current rule qeneration program that sugqests various rules of interpretation for th? ES data summarized by r NTS~JM. Although not finished, the program can provide useful assistance in practical theory formation. f. Ancillary Techniques 1. The !!S facility provides other types of experiments in MS, including ultra-high resolution measurements (masses determined via peak matching), defocussed metastahle ion determinations (Barber-Rlliott technique) and low ionizing voltage experiments. These data are utilized by both scientists and proqrams where anpropriate. 2. Additional computer proqraas provide added problem- solving assistance. 3, Predictor program for predicting major f.. eatures of mass spectra. h. Proqcams for drawinq and displayinq chemical structures. c. Subroutines developed in conjunction with or existing as parts of the Structure Generator for pr2bleol.s of partitioning, construction >f vertex-graphs, and constructive graph labelling. These can be 3nnlie? to answer certain questions of isomerism which do not require the complete generator. For example, the labelling alqorithm can list alI structlrres resulting from substituting sites of a carbocyclic skeleton with stated numbers of different fuactianal qrouns. 9* Other Spectroscopic Techniques ?vailahle to us are the facilities of Professor Djerassi's Laboratory for work requiring Idditional spectroscopic data. Also Ivailahle on a fee for service basis are extensive spectroscopic facilities (NMR, I.!?., 3n.d V.V.) of the chemistry department. These would be utilized for collecting additional data on particular structure problems and gatherinq data on known compounds [particularly in the area of 13CMR) as the AI programs bec?me kn?zledqable about other spectrosconic information. h, Chnaicsl Facilities l-he stsff and facilities of thP chemistry department represent SubstaRt-la1 synthesis capabilities and gerieral chemical know-how. This r?sollrce can be called upon to pr~vid? assistance i.~ 5vnthesi.s of model or labelled cnmpounds, derivatization of 9ixCure9, 3nd so forth, For example, a graduate student in shemistry is presently engaqed in tfiesis research dealing with the lahnratnry synthesis of a now s.strDgen metabolite stronqly su~~3cte3 to be a component of certain preqnancy urines. The prsviously proposed structure 3f this compound was one of the candidate structures inferred by the nlanner in a study of ?strDqen mixtures (Il-dehydroestradiol-17-alpha, ref. 33). 4. !Js?r Community RzDn3rnic utilization of existinq and proposed facilities can be ce3ilized by sharinq them with 1 community of users. Lacking suppfcmentsry funds that would he needed for a comprehensive, na!nr service facility, this =omaunity will include the following 7 r3'105, hot will be informally available t.o others. A. Stanford Eommunity if Stanford Chemistry Department (except for Hodgson, all 3re heavily supported by the hiT!i in their research efforts) Letters of interest are sttached to the proposal in Appendix A. ?rof. C. Djerassi - Steroids, marine sterols Prof. W. Johnson - steroi3s Prof. E. Van Tamelen - steroids, triterpenoids, other natural products Prof. H. Mosher - natural products (e.g., marine toxins) Prof. K, FIodqson - hiDlogical liqands, ligand-metal complexes Prof. J. Collmnr. - cytDchrDme P450 models ii) Stanford ?ledical School Collaborators The following research projects in the Stanford Biomedical ComEunity will furnish samples for mass eztrometric analysis under the present proposal. it)tached to this proposal (Appendix A) are copies of the letters of interest in the proposed facility r?zzived from the principsl investFqators of these arants. Dr. James R. Trndell, Department of Anesthesia, Stanford University School of Yedicinz. Drug metabolite identification in humans. Dr. Irene S. Forrest, biomedical P,esearch Laboratory, Veterans Administration i-I3spital, Palo Alto, Drug tnetabolite identification in bumar.s. Dr. T. Rabinowitz and D.I. Wilkinson, Department of Dermatoloqy, Stanford University School of Medicine. Prostaqlandins. PrDf. Fuqene D. Eiohin, Department of Respiratory Medicine, Stanford Opiversity School of Tedicinc, Ratio of NAD+/WADH in cplls by measuring ratio of oxidized to reduced redox plLrs. Dr. Leo E. Holfister, Veterans Administration 43 Hospital/Department of Medicine, Stanford 1Jniversity School of Fedicine, Mataholism of Marihuana. Dr. Hiram ??. Sern, Pharmacy Deoartment, Stanford University n0spita.l. Drug Identification. Dr. Sumner p1. dalman, Department of Pharmacology, Stanford University School of Medicine. Drug and drug metabolitc identification. Dr. Jack Rarchas, Department of Psychiatry, Stanford University School of Yedicine. Neurotransmitters and- related compounds in m?n. Dr. Keith A. Kvecvolden, Chemical Evolution Branch, YRSA Ames Research Center, Yountain View, Calif. Amino acids, acids in geochemical samples, structure of products formed from electrical discharges in gas mixtures. Dr. William P. Fair, Department of Urology, Stanford University School of Medicine, Identification of the prostatic antibacterial factor; polyamines fputrescine, spermnine, spermidine) in body fluids of patients with prostatic carcinoma, R?sil,?s tha user projects just summarized, other major prospects are in sight, At the time of writing, the chair of pharmacology is vacant. Conversations uith the leadinq candidate have ina icated a deep-seated interest in GC/HR?IS as the principal analytical tool for broad ranging studies of drug metabolism in nal. 3. Rxtramural Users The development of the techniques of OP,D, MS and MCD at Stanford has beer paralleled with extensive sharing of these resources natian- an1 world-wide 513 collaborative research efforts, without anv ~~l~ditional funding. Rather than provide routine service, axneriance has shown th.at discretionary selection of problems results in better utilization 3f our people an,d instrumentation resource;. we would extend this provision of services including atr2ilablp computer programs, to a limited number of extramural IISerT, vote, for example, our successful collaboration with ?r?fassor Adlercreutz, Meilahti Hospital, University of Helsinki, on th? identification of estrogens fron body fluids utilizing the 9X planning proqram {ref. 33). ". Re?atiDnship to ATE-SrJMEY and the "Jenctics Research Center I'he present application is strengthened by two research projects related to, but not overlapping, the proposed research of this scant. 1) RT"P-SUYEX (NIH RR-00785, Oct. 1, 1973, thru July 31, 1978, prirlcipal Investigator, J. Lederberg). This is a resource grant co cst_3tbfi ,sh a national facility for applications of artificial i?tslliqenzn in medicine (AI%). Our own use of this facility will inzlud? SUMEX PDP-10 computer time and file storage necessary to run the DENDRAL artificial intelligence proyrnms. This support vi.11 he furnished without charge to the present proposal. It represents an annual investmerzt of about R100,OOO in computer time z?ui.valsnt value. ?he AT?!-SVEX computinq facility is shared equally between a national user community {AIFP) and a Stanford fledicnl School community. The DENDRAL research will he supported out of the Stanford portion. The ATM service rill be administered under the policg control of a national advisory committee and will be imolemsnted over a national computer network. AIM-SIIFIFX provides tt?p W?3?lS for members of thm national user community interested in ?tr!lzt!lre elucidation to access the DSNDRAL programs. 2) ;enetizs Research Center (NIH POl-SM 20832-01 - approved by thP NISMS Zouncil, awaitinq fundinq, Principal Investiqator, 3. i~33tberq). This research proposal is a comprehensive grant uhich would support interdepartmentni research at the Stanford Eedical 3ch301 in Yedical Senetics, Pe3irltrics and other clinical an3lications. .9 section of that proposal concerns the use of SZ/LRMS for screening body fluids for evidence of inborn errors of metabolism. (This project qre% out of the initial DENDRAL grant, one of the research qoals of nhich was the analysis of body fluids using ?????? o This research on inborn metabolic errors will be n~~?ii!~cted jointly in the Stanford Departments of Genetics and Pediatrics using existing eguipment {Finniqan 1015 Quadrupole mass spectrometer, Varian Aerograph GC and a PDP-11/20 based data ??????? o k?? 3ppr?zisted the value of GC/HPYS analyses of selected extracts 3f ho4y ffjids (i.e., those containinq metaboli tes not identified bv routine GC/LRMS data) when formulating the Senetics fiescarch Lleater proposal. Accordingly, A small amount of funding was there reg~lestad for recording selected GZ/HRYS data on the GC/Varian EAT 711 'Rass spectrometer in the Department of Chemistry. If these funds are awarded, we will noqotlate with YTH a suitable elimination of this minor overlap with the present budget. rr. Sn'?CIFTC AIMS 1'52 spacj.fic aims enumerated in this section will be pursued in tfi3 hiqhly inter-disciplinary manner that has characterized the Ll)7NPRkL project from the start of its NIH support. The aims are not Iisjoint,but interactive and inter-dependent. For example, the pawor of P!S and, potentially, other spactroscopic techniques, z3tn h? enhanced by the use of computer proqrams to perform various asnez+s of structure elucidation and thaory formation, From the star3point of computer science, one measure of the utility of techniques of artifizi al intelligence is how well they perform in r??l-v3rld 3Dplications. I+ is necessary in the development of tFi3s2 proqrams to have a source of data and informed, involved teaB-mates able to criticize methods and results. The aims are alaborsted in the methods section. me have attempt28 tr, keep the proposal to a readable length. rhF?r?fore, some detail. has been omitted. Houever, many details can be found in the bilioqraphy an? we are prepare3 to provide 3dditi onal information ;turinq t.he site visit. 1. Enhance the power of the MS resource. m he existinq MS resource, toqether with computer programs which ixist Dr which are proposed (see Aim 2, below), is capable of -olvinq sane of the structure 3 zlucidat.ion problems of the user community given computer supnort for data collection and re3~lction. We refer specifically to the areas of GC/LRMS and ro:lt%ne, b3tch HRYS samples. We believe that many of the problems of the user community require more powerful techniques (see YPction III) * These techniques, specifically fJC/HRZS and senl-automatic metastable defocussinq, can be provided with a minimum of cost and effort, thus enhancing considerably the capabilities of the resource. Dur first aim is to provide the resource with adequate computer 3upp3rt (re placinq the previous ACZ!E system) to enable collection In3 reduction of mass spectral data including low and high resolution scans and data or defocussed mctastable ions. 83 DrnDoSC to develop this computer support in the ways described 30131, (those aims are writt.en to include the work necessary to im3lencnt the extended PDP-11/20 computer system. A description of the rationale for this choice is provided in Section 1II.A and tha specific auqmentations in the Rudqet Justification). ,!I 1 Convert Pxistinq, proven data acquisition and reduction nc3qra?rs from the PL/ACME langllaqe into Fortran, consistent with t; n3-:: rikizal assembly lanquaqz proqrams for data acquisition snd instrument control. These programs will be written in Fortran to enhance comnatibility with %hc computer systems of other users of 3uch pzlcka~es* 3) Modify these proqrams, 3s requirej, to handle acquisition and ce?3u:tion of frequent or repetitive HRYS scans with selected instLumen~ performance feedhsck to the operator, and to take aqvantaqe of the cxpandcd canabd. A formal proof will be devised =ls W?ll. This s'lqorithm represents one very powerful approach to t hn problem of implementation 3f constraints, 3s discussed in the f~3 ?.ovinq paraqraph. The qencratinq proqrams will be modified to allow isomer qpnQr3tion within constraints, Different kinds of constraints can h? inferred from different kinds of spectroscopic data. !?e intend t7 qive thz proqram knowledqe of 3 variety of these. Thp Planner proqrams that infer constraints from mass sDectron?try data will be hrnadened to include additional kn?wlerIqe shout the spectral behavior of classes of compounds of r"levan~e to thcz NIH-sponsored research of the user community. In 3??i.i+i.on, 83 will add the capability for utilization of izf3rma+t3n about chemical isolation procedures (e.q., one expects scijic and neutral compounds in solvent extraction of acidified bo3v fluins) and relative GC retlntior! times (e.g., to admit the possibility of homaloqous series). w F! ~~303s~ to implement a more qeneral method for ir,fsrrirq the i33ntity of thp mol?sular ion whether or not this 1!3pe1cs explicitly in the spectrum, This information is important f3r the Tuccessful operation of the structure qenerator and the n?3rlrlf?r. 'de want the program to use whatever information is 3v3il3bl2 -Ind not depend, as it ctlrrently does, on having knowledqe 3f the structural slass together with inference rules for that class. Tnt?rfsce routines will he written to mak? it easier for other ;nicntis?s to use tF.ese proqrans, We have to wait for an inker3ctivc system before startLng this: AI!?-SI!MEX will he ideal. rr.p'lf/output routines will he crucial to ~ssy use of the system. u~u3vcc, WC als3 want t3 give users the facility to understand the system's reasoning steps so they can t3k~ 34vantage of it. Tn a4ditirn to makinq the compilter programs available through 4T"-TTJYEX, we w0da like to translate parts of the LISP code into another 13nquaqe - for reasons of both efficiency and ?xp3rtability. We have talked with computer professionals at IBM Pesaaroh Canter about usinq the APL lanquaqe. FORTRAN, ALGOL and PL/l arc other lanquages whose merits for our purposes we will explor?. W? wish to continue a low-level of effort on computer proqrams that interpret other kinds of spectroscopic data. Plannicq proqrams similar to the ?IS Planner could be written for automatic analysis of data fro3 3,ther spectroscopic tc:hniques(e.q., IF, IJV) # as we have illustrated for 13CMP (ref. ??? o Th? structure qenerator's view of chemical strlictlre is topDloqizal and is presentlv unconstrained by bond lengths and angles. 3ecauss stereochemjcal considerations are frequently important in structure el.ucidation, we propose to begin corsidsration of stereochemistry in the structure generation and 2vaLuation processes. !! nrolram with detailed knowledge about information ?htain?Sle from various spectroscopic techniques could be written t3 C?X3Zli!l? a list Df candidate solutions and propose experiments necessary and sufficient to distinquish among them. The program would represet?? an extended Predictor (e.q., ref. 27). Be have a Eir- ,t version of a program that suggests "crucial" metastable pz3ks t.0 he snuqht in order to distinguish among car,didate stru~tutz?s. Work on this proqram will con%inuc at a fow level of 3ct!vity, possibly expandinq into areas other than ES. One topic we will continue to pursue is our collaborative effort uith Dr. ;iLd3 Loew, Geneti.cs Department, on the potential application of molez:rllar orbital theory to prediction of mass spectra (ref. 71). Th33ry 3rnation Proqrams:: The rile formation proqran (RtJLEGF:N) will he extended so that i + car? search a larqer space of rules. Present a priori zo3str3int; on the rule qeneration qive us a search reduction from te?s of millions to a thousand possihl? rules, Even though search heuristics now 3110~ efficient search of these possibilities, we want to be able to deal with much larger spaces efficiently, as wh?? the number of primitive predicates is drastically increased, The PULFGEN proqram will. be modified so th-at complex Eraqmectation and rearranqement processes are manipulated nearly 3s easily 3s simple fraqmentations. The program currently finds frlqmentstion rules invnlvinq 3ne 3r two bonds, possibly followed bv hydroqen miqration, Ir. the case of cyclic systems such as =S~CD~ZCS, however, the proqram must be able to work with sets of . . three or more bonds in some cleavages, Interactive proqrams will. be provided 3~ AIM-SUP!:EX for the inv3stiqat3r to query the rule qeneration program. For example, msnv questions now arise about the proqram steps by which the proqram Tnfer; the rules it suggests as explanations of the re7l~larltins. Why, for example, was some particular rule not considered plausible? New d2it.a will have to be selected in order to test the rules 3nd to differentiate amnnq competinq rules. Be will mite a nroqram that suqqests new experiments (i.e., new data to obtain), deosniinq on the nature of the existing rules, The t?st phase of the theory formation program will be written as an evaluation function of each rule against new data. Tn3;7f3r 3s 3ny new experiments are **crucialtl experiments, the evslu-ation function may merely reject a proposed rule. Mostly, 13wevor , riles will have to be evaluated against new data along 3apv dimensions: frequency, strenqth of evidence, uniqueness, simplicity, and the like. We wish to experiment with the whole theory formation program to ,lztermi3e the critical aspects of our design. For example, 11) how sensitive is the proqram to discrepancies, inconsistencies and ecr:>r5 in the data? (2) how well can the proqram find rules within a sliqhtly different model of chemistry? (3) how well can the proqrsm perform w?%h one pass through the data, or several passes? and (4) how critical are the principles of theory Eormati~on? 3. Apply the structure plucidati3n techniques - both i?;tcumsnt3tion and computer programs - to hFomedically relevant compounds. 3ur own interests are in elucidatinq the structures of, and SnIerstandFnq the MS of, marine sterols, hormonal steroids, and compounds isolated from human bofly fluids that can be associated with qenatic disorders (from research in the GRC). In addition, W? riff be workinq closely with members of the Stanford Medical School 3nd Chemistry Department - in particular those mentioned aih3ve fSect?.on 1.3.4) - on their structure elucidation problems in which YS will be used. Althouqh most users expect to require HRMS 3na SZ/HRYS data, some of their problems will be attacked utilizicq SC/LRMS techniques and library search through (usually) restricted libraries of mass slnectral data, We propose to investiaate some extensions to the technique of library search (SIP Methods) to complement our existinq and planned DENDRAL prcqrams. WE! plan to continue 3ur exchanqe of mass spectral data an? library search information as we have previously done with Dr. 5, Uarkey (University of Colorado Medizsl School) and Dr. F. W. YcLaffnrtv (Cornell Universi.tvI * P.s !.n the Dast, attention to new biomedFcnL research problems will lead t-o increased capabilities in the computer proqrams. We rejuir? close communication with the people engaged in the research so that the proqrams actually assist the researcher while iycreasinq in power, Collahor3 tive proposals have come out of sjxzh past DENDRAL sponsored work, for exampl?, larqe portions of t ?-lP ;9C Droposal apd a proposal. for 13CMH research. k'3 2~visi.92 the interaction and collaboration with the user zoamunity to involve the followinq: