From nmeier@nirvana.imo.physik.uni-muenchen.de Thu Jun 3 04:54:39 1993 From: Christoph.Niedermeier@physik.uni-muenchen.de Message-Id: <9306030840.AA20901@hegel.imo.physik.uni-muenchen.de> Subject: Hierarchical multipole solvers To: chemistry@ccl.net (Chemistry mailing list) Date: Thu, 3 Jun 93 10:40:16 MET DST Hi everybody, I'm working as a PhD student in the field of MD simulations of proteins with special interest in electrostatic interactions. Currently I am developping a method for efficient computation of long range electrostatic interactions in MD simulations of proteins. The most efficient methods existing sofar are, to my knowledge, hierarchical multipole algorithms which scale with O(N log N) (N being the number of atoms in the system). A special variant of this type of algorithms, the socalled Fast Multipole Method (FMM) (Greengard & Rohklin) even scales with O(N). My question to the list is the following: Has anybody worked with this type of algorithms and/or knows of other people who did ? If so, could you please give your experiences with/ opinions on these methods and, if available, supply a list of references ? I will post a summary of responses to the list. Thank you a lot Chris -- Christoph Niedermeier -- Theoretische Biophysik -- Institut fuer medizinische Optik -- Ludwigs-Maximilian-Universitaet Muenchen -- __o Theresienstrasse 37 -- 8000 Muenchen 2 -- Germany _`\<,_ phone: ++49-89/2394-4580, fax: ++49-89/2805248 (_)/ (_) email: Christoph.Niedermeier@Physik.Uni-Muenchen.DE ~~~~~~~~~~~ From scsupham@reading.ac.uk Thu Jun 3 08:49:55 1993 From: scsupham@reading.ac.uk Date: Thu, 3 Jun 93 13:30:01 BST Message-Id: <7375.9306031230@scsscsc3> To: chemistry@ccl.net Subject: Basis sets for transition metals ---------- X-Sun-Data-Type: text X-Sun-Data-Description: text X-Sun-Data-Name: text X-Sun-Content-Lines: 12 Hi, Can anyone help me find good gaussian basis sets for 2nd row transition metals suitable for use with Gaussian 92 ? I`m particularly interested in Mo , Cr, Mn and W if possible. Thanks for any help. john upham John Upham, Dept. of Chemistry, University of Reading, Berks., RG6 2AD, UK. Email: scsupham%reading.ac.uk@uk.ac (BITnet), scsupham@uk.ac.reading (Janet) Voice: +44 734 875123 x7441 (day), Fax: +44 734 311610 ---------- X-Sun-Data-Type: default X-Sun-Data-Description: default X-Sun-Data-Name: .signature X-Sun-Content-Lines: 3 John Upham, Dept. of Chemistry, University of Reading, Berks., RG6 2AD, UK. Email: scsupham%reading.ac.uk@uk.ac (BITnet), scsupham@uk.ac.reading (Janet) Voice: +44 734 875123 x7441 (day), Fax: +44 734 311610 From figuei@lutece.rutgers.edu Thu Jun 3 04:59:04 1993 Date: Thu, 3 Jun 93 08:59:04 -0400 From: figuei@lutece.rutgers.edu (Francisco Figueirido) Message-Id: <9306031259.AA17288@lutece.rutgers.edu> To: Christoph.Niedermeier@physik.uni-muenchen.de Subject: Hierarchical multipole solvers I am currently testing an implementation of the Barnes-Hut algorithm that I have inserted into a MD simulation package (IMPACT) we use here at Rutgers. I haven't tried the FMM, though (it is known to have a much larger overhead than the O(N log N) method). These are the references I have collected so far (in BibTeX format): @Article{appel, author = "Andrew W. Appel", title = "An efficient program for many body simulations", journal = siamscistat, volume = "6", pages = "85--103", year = "1985", abstract = "The simulation of $N$ particles interacting in a gravitational force field is useful in astrophysics, but such simulations become costly for large $N$. Representing the universe as a tree structure with the particles at the leaves and internal nodes labeled with the centers of mass of their descendants allows several simultaneous attacks on the computation time required by the problem. These approaches range from algorithmic changes (replacing an $O(N^2)$ algorithm with an algorithm whose time-complexity is believed to be $O(N\log N)$) to data structure modifications, code-tuning, and hardware modifications. The changes reduced the running time of a large problem ($N=10000$) by a factor of four hundred. This paper describes both the particular program and the methodology underlying such speedups.", } @Article{barnes:hut, author = "Josh Barnes and Piet Hut", title = "A hierarchical ${O}({N}\log {N})$ force-calculation algorithm", journal = "Nature", volume = "324", pages = "446--449", year = "1986", abstract = "Until recently the gravitational $N$-body problem has been modelled numerically either by direct integration, in which the computation needed increases as $N^2$, or by an iterative potential method in which the number of operations grows as $N\,\log N$. Here we describe a novel method of directly calculating the forces on $N$ bodies that grows only as $N\,\log N$. The technique uses a tree-structured hierarchical subdivision of space into cubic cells, each of which is recursively divided into eight subcells whenever more than one particle is found to occupy the same cell. This tree is constructed anew at every time step, avoiding ambiguity and tangling. Advantages over potential-solving codes are: accurate local interactions; freedeom from geometrical assumptions and restrictions; and applicability to a wide class of systems, including (proto-)planetary, stellar, galactic and cosmological ones. Advantages over previous hierarchical tree-codes include simplicity and the possibility of rigorous analysis of error. Although we concentrate here on stellar dynamical applications, our techniques of efficiently handling a large number of long-range interactions and concentrating computational effort where most needed have potential applications in other areas of astrophysics as well." } @Book{greengard, author = "Leslie Greengard", title = "The Rapid Evaluation of Potential Fields in Particle Systems", publisher = "The MIT Press", address = "Cambridge, Massachusetts", year = "1988", } @PhDThesis{draghicescu, author = "Draghicescu, Cristina I.", title = "Efficient Algorithms for Particle Methods", school = "The Pennsylvania State University", year = "1991", abstract = "A fast algorithm is presented, which reduces the amount of work necessary for computing pairwise interactions in a system of $n$ particles from $O(n^2)$ to $O(n(\log n)^p)$, where $p$ depends on the problem in question. Error and work estimates are given.\par I illustrate its application to the approximation of the Euler equations in fluid dynamical simulations using the point vertex method. The algorithm can be applied for both two- and three-dimensional simulations; in the first case I show that, with a proper choice of parameters, the accuracy and stability of the direct method are preserved.\par Also discussed is the application of the algorithm to the problem of evaluating interactions in molecular simulations. A slightly modified version can be used to reduce the complexity of the integral equation method for boundary value problmes. I implemented the algorithm for such a problem and provide the numerical results. On a SUN 4 the algorithm reduces the CPU time required for a calculation with 500,000 points from a month to 15 minutes and is three times faster than the direct method for as few as 128 particles.", } @PhDThesis{salmon, author = "John K. Salmon", title = "Parallel Hierarchical ${N}$-body Methods", school = "California Institute of Technology", year = "1991", abstract = "Recent algorithmic advances utilizing hierarchical data structures have resulted in a dramatic reduction in the time required for computer simulation of $N$-body systems with long-range interactions. Computations which required $O(N^2)$ operations can now be done in $O(N\,\log N)$ or $O(N)$. We review these tree methods and find that they may be distinguished based on a few simple features. \par The Barnes-Hut (BH) algorithm has received a great deal of attention, and is the subject of the remainder of the dissertation. We present a generalization of the BH tree and analyze the statistical properties of such trees in detail. We also consider the expected number of operations entailed by an execution of the BH algorithm. We find an optimal number for $m$, the maximum number of bodies in a terminal cell, and confirm that the number of operations is $O(N\,\log N)$, even if the distribution of bodies is not uniform. \par The mathematical basis of all hierarchical methods is the multipole approximation. We discuss multipole approximations, for the case of arbitrary, spherically symmetric, and Newtonian Green's functions. We describe methods for computing multipoles and evaluating multipole approximations in each of these cases, emphasizing the tradeoff between generality and algorithmic complexity. \par $N$-body simulations in computational astrophysics can require $10^6$ or even more bodies. Algorithmic advances are not sufficient, in and of themselves, to make computations of this size feasible. Parallel computation offers, {\em a priori\/}, the necessary computational power in terms of speed and memory. We show how the BH algorithm can be adapted to execute in parallel. We use orthogonal recursive bisection to partition space. The logical communication structure that emerges is that of a hypercube. A local version of the BH tree is constructed in each processor by iteratively exchanging data along each edge of the logical hypercube. We obtain speedups in excess of 380 on a 512 processor system for simulations of galaxy mergers with 180,000 bodies. We analyze the performance of the parallel version of the algorithm and find that the overhead is due primarily to interprocessor synchronization delays and redundant computation. Communication is not a significant factor." } @Article{SL:JStatPhys:91, author = "K. E. Schmidt and Michael A. Lee", title = "Implementing the Fast Multipole Method in Three Dimensions", journal = "J. Stat.\ Phys.{}", volume = "63", pages = "1223--1235", year = "1991", abstract = "The Rokhlin-Greengard fast multipole algorithm for evaluating Coulomb and multipole potentials has been implemented and analyzed in three dimensions. The implementation is presented for bounded charged systems and systems with periodic boundary conditions. The results include timings and error characterizations.", } @Article{Hernquist:JCP:87, author = "Lars Hernquist", title = "Vectorization of Tree Traversals", journal = jcompphys, volume = "87", pages = "137--147", year = "1990", abstract = "A simple method for vectorizing tree searches, which operates by processing all relevant nodes at the same depth in the tree simultaneously, is described. This procedure appears to be general, assuming that gather-scatter oprations are vectorizable, but is most efficient if the traversals proceed monotonically from the root to the leaves, or {\em vice versa\/}. Particular application is made to the hierarchical tree approach for computing the self-consistent interaction of $N$ bodies. It is demonstrated that full vectorization of the requisite tree searches is feasible, resulting in a factor $\approx$ 4--5 improvement in cpu efficiency in the traversals on a CRAY X-MP. The overall gain in the case of the Barnes-Hut tree code algorithm is a factor $\approx$ 2--3, implying a net speed-up of $\approx$ 400-500 on a CRAY X-MP over a VAX 11/780 or SUN 3/50.", } @Article{Makino:JCP:87, author = "Junichiro Makino", title = "Vectorization of a Treecode", journal = jcompphys, volume = "87", pages = "1990", year = "148--160", abstract = "Vectorized algorithms for the force calculation and tree construction in the Barnes-Hut tree algorithm are described. The basic idea for the vectorization of the force calculation is to vectorize the tree traversal across particles, so that all particles in the system traverse the tree simultaneously. The tree construction algorithm also makes use of the fact that particles can be treated in parallel. Thus these algorithms take advantage of the internal parallelism in the $N$-body system and the tree algorithm most effectively. As a natural result, these algorithms can be used on a wide range of vector/parallel architectures, including current supercomputers and highly parallel architectures such as the Connection Machine. The vectorized code runs about five times faster than the non-vector code on a Cyber 205 for an $N$-body system with $N=8192$.", } @Article{Barnes:JCP:87, author = "Joshua E. Barnes", title = "A Modified Tree Code: Don't Laugh; It Runs", journal = jcompphys, volume = "87", pages = "161--170", year = "1990", abstract = "I describe a modification of the Barnes-Hut tree algorithm together with a series of numerical tests of this method. The basic idea is to improve the performance of the code on heavily vector-oriented machines such as the Cyber 205 by exploiting the fact that nearby particles tend to have very similar interaction lists. By building an interaction list good everywhere within a cell containing a modest number of particles and reusing this interaction list for each particle in the cell in turn, the balance of computation can be shifted >from recursive descent to force summation. Instead of vectorizing tree descent, this scheme simply avoids it in favor of force summation, which is quite easy to vectorize. A welcome side-effect of this modification is that the force calculation, which now treats a larger fraction of the local interactions exactly, is significantly more accurate that the unmodified method.", } @Article{Makino:JCP:88, author = "Junichiro Makino", title = "Comparison of Two Different Tree Algorithms", journal = jcompphys, volume = "88", pages = "393--408", year = "1990", abstract = "The efficiency of two different algorithms of hierarchical force calculation is discussed. Both algorithms utilize the tree structure to reduce the cost of the force calculation from $O(N^2)$ to $O(N\log N)$. The only difference lies in the method of the construction of the tree. One algorithm uses the oct-tree, which is the recursive division of a cube into eight subcubes. The other method makes the tree by repeatedly replacing a mutually nearest pair in the system by a super-particle. Numerical experiments showed that the cost of the force calculation using these two schemes is quite similar for the same relative accuracy of the obtained force. The construction of the mutual-nearest-neighbor tree is more expensive than the construction of the oct-tree roughly by a factor of 10. On the conventional mainframes this difference is not important because the cost of the tree construction is only a small fraction of the total calculation cost. On vector processors, the oct-tree scheme is currently faster because the tree construction is relatively more expensive on the vector processors.", } From jstewart%fujitsu@uunet.UU.NET Thu Jun 3 00:20:35 1993 Date: Thu, 3 Jun 93 07:20:35 PDT From: jstewart@fai.com (Dr. James Stewart) Message-Id: <9306031420.AA22214@fujitsu.fai.com> To: chemistry@ccl.net Subject: MOPAC 93 MOPAC 93 MOPAC 93, the next version of MOPAC after MOPAC 6, is now available through the Quantum Chemistry Program Exchange. To get a copy, contact Richard Counts Quantum Chemistry Program Exchange Creative Arts Building 181 Indiana University Bloomington, IN, 47405 Tel (812) 855-4784 Fax (812) 855-5539 E-mail COUNTSR@IUBACS Funding for the continuing development of MOPAC has been provided by Fujitsu Limited. In order to control the distribution and development of MOPAC, and because Fujitsu is a commercial organization, MOPAC 93 will be protected by copyright. In recognition of the public domain status of earlier MOPACs, end users can obtain a site-license copy of MOPAC 93 from QCPE at the standard QCPE cost. A special license must be obtained before MOPAC 93 can be re-distributed. Interested parties should contact Miss Kumiko Iguchi Senior Application Analyst Fujitsu Systems Business of America, Inc., 5200 Patrick Henry Drive, Santa Clara, CA 95054 Tel: (408) 988 8012 X 226 FAX: (408) 492 1982. MOPAC 93 was developed on a SUN SPARC 2 workstation, and the test data results included with MOPAC 93 can only be duplicated on a SUN SPARC 2, due to timing considerations. However, I expect that users will have little or no difficulty in getting MOPAC 93 to run on any scalar UNIX platform. James J. P. Stewart Tel: (719) 488-9416 FAX: (719) 488-9758 Email: jstewart@fai.com From jle@world.std.com Thu Jun 3 06:48:56 1993 Date: Thu, 3 Jun 1993 10:48:56 -0400 From: jle@world.std.com (Joe M Leonard) Message-Id: <199306031448.AA02893@world.std.com> To: chemistry@ccl.net Subject: Thanks for the info! Well... I got an amazing number of replies for my recent post re: 8-bit color X for the Mac, and the consensus (by far) was to use MacX. The only other package mentioned was White Pines eXodus package, but folks weren't real up on the current status of this product. I'll follow up if there's any unusual problems with trying to use them, but I don't expect any... Joe Leonard jle@world.std.com From fagerbur@Kodak.COM Thu Jun 3 06:47:34 1993 Date: Thu, 3 Jun 93 10:47:34 -0400 Message-Id: <9306031447.AA04481@Kodak.COM> From: fagerbur@Kodak.COM To: "chemistry@ccl.net"@Kodak.COM Subject: Zeolite summary Hi Netters! I apologize for the delay in sending this summary. First, many thanks to all who responded to my query on zeolites. There is obviously a lot of info out there to start out with. I enclose the summary hoping that I have removed at least most of the duplication. There were a number of messages that said that I out to contact Biosym as they have a consortium that has looked into this very thing (at least the zeolite structures themselves). A talk with those guys has revealed that they are about ready to come out with some parameters, etc. in the next little while. It sounded to me like they had done a really thorough job so it will be interesting to see the package especially since they seem to be really good at documenting what they do and why. John Newsom suggested the following leading references: Modeling of Structure and Reactivity in Zeolites C. R. A. Catlow Ed. Academic Press, London (1992) R. A. Jackson and C. R. A. Catlow Molecular Simulation 1 pp. (1988) J. B. Nicholas, A. J. Hopfinger, L. E. Iton and F. R. Trouw J. Amer. Chem. Soc. 113 4792 (1991) G. J. Kramer, N. P. Farragher, B. W. H. van Beest and R. A. van Santen Phys. Rev. B 43 5068 (1991) M. W. Deem and J. M. Newsam J. Amer. Chem. Soc. 114 7198-7207 (1992) There are further improved silicate and aluminosilicate potentials for zeolite frameworks that do a reasonable job on both structure and vibrational spectra:- J. Hill Ph. D. Thesis, Humboldt University, Berlin (1992), but these results have not appeared in literature publications as yet. If you are interested in more information on same please let me know. On the organics interacting with zeolites front you might start with the first reference again and perhaps (with citations that provide access to the earlier work)- J. B. Nicholas, F. R. Trouw, J. E. Mertz, L. E. Iton and A. J. Hopfinger J. Phys. Chem. 97 4149-4163 (1993) R. L. June, A. T. Bell and D. N. Theodorou J. Phys. Chem. (1990) There are quite a number of studies here, but the references in the above provide a reasonable entry. Other useful starts are the various conference proceedings volumes, particularly of the Int. Zeolite. Association. Luis Montero wrote: The Laboratoire de Dynamique des Interactions Moleculaires (CNRS) in the Uni- versity of Paris VI is carrying out a nice work in theoretical modeling of zeolites. We are working with them from Havana (my own Lab. of Computational and Theoretical Chemistry, in the University of Havana, and the Institute of Pedagogy there). You can follow papers from the last J. Phys. Chem. 1993, 97, 641 and references therein. I particularly recomend you to address Dr. Earl Evleth (udim018@frors31.bitnet) who is very experienced. He is now in US in vacances, but returning to Paris in one or two weeks. >From Adi M. Treasurywala: Look at the papers fo Bastian Van De Graff, Julianalaan 136 2628BL Delft The Netherlands I have been listening to lectures by him about this topic for about ten years now. He seems to me to be doing excellent work in this area. If he himself has not published anything (a TRUELY REMOTE possibility) then he will surely be able to put you onto the relevant literature. David Edelson wrote: Try Computers & Chemistry, the article by Leherte et al, Vol 15 p 273 (1991) and extensive references in it. Michael Sennett suggested: starting point.... Gubbins et al., J. Phys. Chem., 1993, vol. 97, p. 494 A book with potential relevance.... Computer Simulation and the Statistical Mechanics of Adsorption, aD. Nicholson and N.G. Parsonage, Academic Press, Oxford, 1982. A reference to MD simulations.... Sokolowski and Fischer, Mol. Phys., 1990, vol. 71, p. 393 A paper treating surface characteristics of metal oxides... Salasco et al., Mol. Phys., 1991, vol. 72, p. 267 i) A force field for all-silica zeolites was developed by E. de vos Burchardt et al Zeolites 1992 12 183 ii) Biosym include zeolite parameters within their DISCOVER force field - used for hydrocarbon/zeolite interaction studies - Freeman et al Chem Phys Letters 1991 186 137 iii) Mabilia et al have developed parameters ( bending ) for the modelling of the sodalite cage JACS 1987 109 26 iv) Titiloye et al have produced parameters for the interaction of hydrocarbons with zeolite frameworks J Phys Chem 1991 10 4039 v) Kramer et al have a parameter set derived from ab initio data Phys Rev B 1991 43 6 5068 However, the chances are that you are interested in the interaction of host molecules within the zeolites - in which case, decent parameters are very much harder to find - these may have to be developed specifically for the study you are undertaking. A good review on ab initio calculations on zeolites ( If you are interested ) is Sauer J Chem Rev 1989 89 199. Information from Dave Doherty: I only have pre-prints of these, but John Nicholas and Tony Hopfinger did a series of papers on this subject over the last couple of years. I don't have the actual references handy, but I they appeared in JACS. John did a *very* careful job of parameterizing a zeolite force field, and then did a series of simulations with various small organics bouncing around in cages to study diffusion. Info from Randy Snurr: "Introduction to Zeolite Science and Practice" Ed. by H. van Bekkum, E.M. Flanigen, and J.C. Jansen Chapter 9 "Introduction to Zeolite Theory and Modelling" by R.A. van Santen, D.P. De Bruyn, C.J.J. den Ouden, and B. Smit This chapter covers both areas and has lots of references. On simulations of zeolite structures (e.g. examining relative stability), I would suggest looking for papers by R.A. van Santen C.R.A. Catlow J. Sauer (quantum calculations) For simulations of sorbates inside zeolites, I could recommend the following from our group here at Berkeley: R.L. June, A.T. Bell, D.N. Theodorou "Prediction of low occupancy sorption of alkanes in silicalite," J.Phys.Chem. 1990, 94, 1508. "Molecular dynamics study of methane and xenon in silicalite" J.Phys.Chem. 1990, 94, 8232. "Transition-state studies of xenon and SF_6 diffusion in silicalite" J.Phys.Chem. 1991, 95, 8866. "Molecular dynamics studies of butane and hexane in silicalite" J.Phys.Chem. 1992, 96, 1051. R.Q. Snurr, R.J. June, A.T. Bell, D.N. Theodorou "Molecular simulations of methane adsorption in silicalite" Molec. Simulation 1991, 8, 73. Brian Teppen information: 1992: J. Chem. Soc., Chem. Commun. 1992:879 J. Am. Chem. Soc. 114:7198 Chem. Phys. Lett. 188:320 1991: J. Phys. Chem. 95:4038 J. Am. Chem. Soc. 113:6435 J. Chem. Soc., Faraday Trans. 87(13):1947-1970 (three articles). Again, my thanks to all who responded. Dave Fagerburg Eastman Chemical Co. Kingsport, TN fagerbur@kodak.com From vasz@bif.creighton.edu Fri Jun 4 06:13:41 1993 Date: Fri, 4 Jun 93 11:13:41 -0500 From: vasz@bif.creighton.edu (Sandor Lovas) Message-Id: <9306041613.AA01639@bif.creighton.edu> To: chemistry@ccl.net Subject: Re:ASED At my best knowledge the extended Huckel MO method modified recently by Calzaferri (J.Phys.Chem. 93,5366) does the same job as ASED. The program called ICON8 is available from QCPE. Sincerely, Sandor Lovas Creighton University Omaha, NE From mail Thu Jun 3 11:47:02 1993 Date: Thu, 3 Jun 1993 12:10:39 -0400 From: hyper!hurst (Graham Hurst) Message-Id: <9306031610.AA23586@hyper.hyper.com> To: Mr Andrew D Allen , chemistry@ccl.net Subject: Re: Electron density maps on a PC Andy Allen writes: > Dear All, > This is deliberatly left general as I haven't got the full details of > what a friend of mine is looking for. The basics are the display of charge or > "electron density" maps on a PC. If it turns out that your friend is looking for a program to calculate and display electron density maps on a PC, HyperChem can do it. (It requires at least a 386 with fpu running Windows in 386 enhanced mode.) Hope this helps, Graham ------------ Graham Hurst Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 internet: hurst@hyper.com From ZAHAREVITZ@dtpvx2.ncifcrf.gov Mon Jun 3 09:01:00 1993 Message-Id: <199306031805.AA21512@oscsunb.ccl.net> Date: 3 Jun 93 14:01:00 EST From: "Zaharevitz, Daniel" Subject: Theoretical estimation of octanol-water partition coefficient To: "chemistry" Does anyone have any experience with theoretical methods for estimating octanol-water partition coefficients? We are especially interested in methods that might reasonably be appilied to large numbers of structures. Any suggestions or references would be appreciated. Thanks, Dan Zaharevitz zaharevitz@dtpvx2.ncifcrf.gov Information Technology Branch Developmental Therapeutics Program National Cancer Institute From burger@violet.berkeley.edu Thu Jun 3 05:05:59 1993 Date: Thu, 3 Jun 93 12:05:59 -0700 From: burger@violet.berkeley.edu Message-Id: <9306031905.AA23534@violet.berkeley.edu> To: chemistry@ccl.net Subject: M. Whitbeck's kinetic simulation program Dear Netters, I would like to hear from people who have succeded to port Mike Whitbecks kinetic simulation program (available from www.ccl.net) to the MS-DOS platform. Thx in advance Peter From hogue@canada.den.mmc.com Thu Jun 3 08:02:12 1993 Date: Thu, 3 Jun 93 14:02:12 MDT From: hogue@canada.den.mmc.com (Pat Hogue 1-2183) Message-Id: <9306032002.AA00572@canada.den.mmc.com> To: chemistry@ccl.net Subject: Shortage of Computational Chemists? I am just finishing a masters degree using computational techniques. This new found skill has not made any difference to prospective engineering or chmical employers. Thank God that the the threatened layoff has not occurred! I don't know if my aerospace background is the problem or if it is the oversupply problem. Pat Hogue From jaeric@mtcamm.monsanto.com Thu Jun 3 10:38:03 1993 From: Jon A. Erickson Message-Id: <9306032038.AA04978@mtcamm.monsanto.com> Subject: UV-VIS simulation responses... To: chemistry@ccl.net (OSU Comp. Chem. List) Date: Thu, 3 Jun 93 15:38:03 CDT Here is a summary of the responses to the following posting: Does anyone know of any work involving the simulation/prediction of UV-Vis spectra? Any code or references would be of help. Thanks to those who responded. ################################################################ # Jon Erickson e-mail: jaeric@mtcamm.monsanto.com # # Monsanto Company, U3E phone: (314) 694-1511 # # 800 N. Lindbergh Blvd. # # St. Louis MO, 63167 # # # # What is a magician but a practising theorist? # # -- Obi-Wan Kenobi # ################################################################ ------------------------------------------------------------ From: scsupham@reading.ac.uk Dear Jon, I would like to see the replies to your question. I suspect most work has been done with the INDO/S algorithm (part of ZINDO) using CI methods. I assume you are aware of ZINDO. CaChe includes a nice interface to this. john upham John Upham, Dept. of Chemistry, University of Reading, Berks., RG6 2AD, UK. Email: scsupham%reading.ac.uk@uk.ac (BITnet), scsupham@uk.ac.reading (Janet) Voice: +44 734 875123 x7441 (day), Fax: +44 734 311610 ------------------------------------------------------------ From: Paul A Cahill Jon, We use a version of ZINDO available through CAChe scientific that does a fair to excellent job of predicting spectra. ZINDO was developed by Mike Zerner at the Quantum Theory Project at U of Fl, Gainesville. He may be able to send you a version, but I think the input data format is murder. CAChe's version eliminates that difficulty completely. Their ZINDO runs on either an internal 88100-RISC card in a MacIntosh or on a RS/6000. What does good to excellent mean? For most dyes (I work on fluorescent and nonlinear optical materials), it means +/-10% most of the time. Solvent effects can sometimes be taken into account. There are some real failures, however, so one must be careful. Good luck. You might reach CAChe at 1-800-544-6634 (I think) for more info. Paul Cahill Senior Member of Tech Staff Dept 1811 Chemistry of Organic Materials Sandia National Labs Albuquerque, NM 87185 pacahil@somnet.sandia.gov 505-844-5754 505-844-9624 fax ------------------------------------------------------------ From: mxm@biosym.com (Max Muir ) I think the best program for this is ZINDO from Mike Zerner at the University of Florida, Gainesville. His email address is zerner@qtp.ufl.edu. The program is available commercially from the CAChe group, who have written a first class interface to it. It is also available from my own company, BIOSYM. Max Muir PS. If you are interested I could send (gratis) a copy of the documentation for the program to let you see if it is what you're after. ------------------------------------------------------------ From: U6124673@ucsvc.ucs.unimelb.edu.au Dear Jon, For ab initio calculations of UV-VIS spectra you should see Roos, Andersson & Fulsher, Chem.Phys.Letts. 192 (1992) 5. (+refs therein and a futher paper which came out quite recently) Their MCSCF program is available by writing to the authors. Note that good ab initio calcs are expensive and difficult, (eg see McMurchie & Davidson J.Chem.Phys., 67 (1977) 5613). You may also like to look at the CI-singles method in Gaussian-90: Foresman, Head-Gordon, Pople & Frisch, J.Phys.Chem. 96 (1992) 135. Note, however, that it is difficult to get excitation energies close to experiment, however the CIS method is very handy if you're not too worried about that. There are some good semiempirical programs around, eg CNDOS/CI is often used for calculations of photoinduced electron transfer stuff: see eg. Bragg, Broo & Larsson, Chem.Phys., 156 (1991) 1. There are also some INDOS/CI programs around (I'll find the refs if you are interested). Cheers Greg (university of melbourne) e-mail: u6124673@ucsvc.ucs.unimelb.edu.au ------------------------------------------------------------ From: jas@medinah.atc.ucarb.com (Jack Smith) Subject: Re: UV-Vis simulation... Status: RO The ZINDO program is particularly parameterized to do this, especially for inorganic complexes. The CAChe implementation actually constructs a UV-Vis spectra for you from the output, and it will help identify what transitions give rise to specific peaks. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= JACK A. SMITH ...................................................................... Union Carbide Corp. || Phone: (304) 747-5797 Catalyst Skill Center || FAX: (304) 747-5571 P.O. Box 8361 || S. Charleston, WV 25303 || Internet: jas@medinah.atc.ucarb.com -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= ------------------------------------------------------------ From: Andy Holder Dear Dr. Erickson, AMPAC 4.5 contains an extensive set of tools for computing the properties of molecules including a module to estimate UV/Vis absorption bands, both position and relative intensity. We would be happy to do a test calculation or two for you. I have you on my mailing list, but would also be happy to forward some printed matter if you wanted it. Andy =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= DR. ANDREW HOLDER Assistant Professor of Computational/Organic Chemistry Department of Chemistry || BITNET Addr: AHOLDER@UMKCVAX1 University of Missouri - Kansas City || Internet Addr: aholder@vax1.umkc.edu Spencer Chemistry, Room 315 || Phone Number: (816) 235-2293 Kansas City, Missouri 64110 || FAX Number: (816) 235-1717 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= ------------------------------------------------------------ From: David Heisterberg Prof. Bartlett from U. Florida was just here giving a seminar: "Electronic Excitation Spectra: Ethylene to Norbornadiene with Coupled-Cluster Methods" using, of course, their ACES2 system. So you might want to check for some recent papers by Bartlett and Stanton. Dave Heisterberg ------------------------------------------------------------ From: John P Lafemina Jon, for what kind of systems? for organic molecules and polymers (especially aromatic systems) the CNDO/S3 model developed by Charlie Duke and co-workers in the mid-70's has worked and continues to work very well. john lafemina d3c876@pnlg.pnl.gov ------------------------------------------------------------ HyperChem from AutoDesk also has ZINDO, I believe. From jaeric@mtcamm.monsanto.com Thu Jun 3 10:46:21 1993 From: Jon A. Erickson Message-Id: <9306032046.AA05201@mtcamm.monsanto.com> Subject: summary of pdb motif searching.... To: chemistry@ccl.net (OSU Comp. Chem. List) Date: Thu, 3 Jun 93 15:46:21 CDT Here is a summary of responses to the following posting: I am looking for tools/references/methods/any thoughts for finding all occurrences of an arbitrary structural specification (eg. a certain loop in relation to a particular helix) in a protein. Many thanks for all the responses. ################################################################ # Jon Erickson e-mail: jaeric@mtcamm.monsanto.com # # Monsanto Company, U3E phone: (314) 694-1511 # # 800 N. Lindbergh Blvd. # # St. Louis MO, 63167 # # # # What is a magician but a practising theorist? # # -- Obi-Wan Kenobi # ################################################################ ------------------------------------------------------------------- From: Max Hello, I know of at two commercial packages that probably do what you want. One is IDITIS, sold by Oxford Molecular and originally developed at Janet Thornton's lab. The other is PROTE, sold by your TRIPOS neighbors, and origi- nally developed by Peter Willet (a recent, 1992 or 1993, paper in J. Mol. Biol. describes this program). Also, I have seen descriptions of a program called WHATIF by Vriend (I believe at the EMBL in Heidelberg) that may include this type of functionality. I also think Vriend makes this program available (for a fee). I saw a demo of IDITIS by the graduate student who worked on it (and who is now with Oxford Molecular); it looks good. I also worked a little bit with an "alpha" version of PROTE: it is much faster for what it does, although the range of applications is not as wide as that of IDITIS. The bottom line: I think there are a few options out there to do what you want (if you are willing to spend a non-trivial amount of money). I would be interested to find out of any cheaper alternatives your query to the net may produce. Best wishes, ======================================== Max Vasquez Protein Design Labs, Inc. 2375 Garcia Ave. Mountain View CA 94043 Phone (415) 903 3744 FAX (415) 903 3730 e-Mail wk01189 @ worldlink.com ======================================== ------------------------------------------------------------------- From: stoutepf@lldmpc.dnet.dupont.com (Pieter Stouten, +1-302-695-3515) >In response to Jon Erickson's enquiry, I would like to suggest the >program "IDITIS". IDITIS is a relational database for searching >protein structures from the Brookhaven Protein Databank. > I am not sure that a strictly relational approach is very effective when searching protein sequence and structure databases since one often is interested in properties of stretches of residues rather than individual residues. In a purely relational system this requires a large number of operations or columns. Permit me to suggest an alternative program. The multi-purpose modelling program WHAT IF incorporates a very fast and flexible database system (SCAN3D), which uses the relational paradigm whenever possible, but also allows for matching of a stretch of residues with a single comparison. For more information, please contact WHATIF's author Gert Vriend directly. He can be reached at vriend@embl-heidelberg.de. Relevant references are: * G. Vriend, "WHAT IF: a molecular modeling and drug design program", J. Mol. Graph., 8, 5255 (1990). * P.F.W. Stouten & G. Vriend, "How to Extract Non-Trivial Information from Protein Structure and Sequence Data: The SCAN3D Retrieval System," Chem. Design Autom. News 7(5) (1992) 18-23. * G. Vriend, C. Sander & P.F.W. Stouten, "A Property-Profile-Oriented Protein Sequence-Structure Database." Submitted for publication. Hope this helps. Cheers, Pieter Stouten The Du Pont Merck Pharmaceutical Company P.O. Box 80353 Wilmington, DE 19880-0353 Phone + 1 - 302 - 695 - 3515 Fax + 1 - 302 - 695 - 2813 E-mail stoutepf@lldmpc.dnet.dupont.com ------------------------------------------------------------------- From: tripos!david@uunet.UU.NET (David Mosenkis) Jon Erickson, The PROTEP product from Tripos will find all occurences of a specified structural relationship between alpha helices and beta strands in the Brookhaven PDB, or alternatively, occurences of a subset of the pattern of these secondary structural elements. However, it currently does not know about anything except helices and strands. Sybyl also has facilities for doing fast queries for a restricted set of structural criteria from the Brookhaven PDB, including finding fragments of a given length with a given end-to-end distance, and wild-card patterns of sequences (using Unix regular expressions). I would be interested to hear what other responses you get. David Mosenkis (david@tripos.com) Tripos Associates 215-844-3479 ------------------------------------------------------------------- From: rwoods@biop.ox.ac.uk Dear Netters, In response to Jon Erickson's enquiry, I would like to suggest the program "IDITIS". IDITIS is a relational database for searching protein structures from the Brookhaven Protein Databank. I have not used the program yet myself, but others in the group use it with apparent success. Iditis is distributed by Oxford Molecular Ltd. Their number is (from the U.S.) 011-44-865-784600. Good searching, Rob Woods Robert J. Woods, Ph.D. Glycobiology Institute University of Oxford South Parks Road Oxford, OX1 3QU ------------------------------------------------------------------- From: Johan Postma Dear Jon Erickson, For your information: we are working on a long-term project to design "simple" tools to process PDB - data and a "general" query program. Enclodes is a sample program (VMS-FORTRAN/PASCAL/C), showing the functionality of the PDB-access-functions. Additionaly the same "hard-wired" query can be done using a "query-description" file, which is then processed by QUE (general query program). The programs are still in a stage of development. These data-access-functions do not "act" upon the PDB-ASCII files but a "binary" representation of them (produced by a program call YABP). Once this "data-file" has been created any item can be access with high speed and efficient use of memory-resources etc.. The timings of the example programs were done on several VAX's (PDB-binary-file contains (test-case) 334 PDB-Files) VAX 900-420 1 CPU second Micro-VAX 3100.80 2 CPU seconds VAX-station 7 CPU seconds regards. Johan P.M. Postma EMBL Heidelberg Germany ------------------------------------------------------------------- From youkha@iris75.biosym.com Thu Jun 3 10:55:52 1993 Date: Thu, 3 Jun 93 17:55:52 -0700 From: youkha@biosym.com (Philippe Youkharibache) Message-Id: <9306040055.AA25457@iris75.biosym.com> To: chemistry@ccl.net Subject: summary: shortage of compuational chemists Dear netters: In my original posting, I made the following answer/remarks to a question on the net: > >Question for everyone? Is there really a shortage of computational > >chemists/chemists in general? > >And my answer is : there is an incredible shortage of computational >chemists in the chemical industry, all over the world. In fact, it is >much worse in Europe and Japan. (So it is not only a green card issue) > >Many people even wonder HOW one becomes a computational chemist in the >pharmaceutical industry. Was one trained for that ? Some experimental >chemists even wonder: what is computational chemistry ? It certainly would >be interesting to hear about ways one can become a computational chemist >today for many young graduate students. The only way I know is to >come from a lab that does some sort of molecular modeling or theoretical >chemistry , but usually the training one will get there will be narrower >than what will be needed to tackle drug design or protein engineering >problems for example, from a computation standpoint. It certainly will >be a start though. > >Maybe a discussion on what should be the necessary training in academia >or engineering schools for computational chemistry would be of interest; >and how to make that field more central (should I also say reliable ?) in >industry. I have received quite a bit of mail about this posting, mostly questions >from graduate students and young Ph.Ds. who would like to know how to "break-into" computational chemistry, as they put it (more than one used that word. It must be significant.). Questions on how to write a resume to pass screening when looking for an industrial job, and when you have a purely theoretical background. Questions on where are the "job openings" that I am talking about. Some requested background references for computational chemistry, to learn about applications. Most questions came from scientists with a theoretical background either in theoretical (quantum) chemistry, applied math and computer science who would like to "switch" to molecular modeling. First of all, I must apologize to those who might have been mislead by my remarks, and are looking for a job in the immediate future. In saying that there is a "shortage of computational chemists" in the chemical industry, I did not mean to say that there are a lot of job openings right now. An obvious reason is the coming (?) health care system reform and the uncertainty it creates for pharmaceutical companies. I just had in mind to stimulate a discussion on the role that computational chemists and computational chemistry plays in industry, and what training should be required for those who would like to "break into" that field. For those who asked what is industry looking for, just ask yourself what are these companies products, and you'll have your answer in broad terms. A drug company hiring a computational chemist will expect him/her to use state of the art computations to help designing drugs and/or protein structures (or other biopolymers) and properties. Knowing about SOFTWARE used in industry will tell you what computational techniques are used. Stop by these companies booths at shows to learn about them. Ask them some litterature on products, look at their references. These programs are used also in academia (a lot of them come from academia), get access to them. The knowledge of protein structure should be considered as a must, if it is not already. For those looking for references, a very good primer is Dickerson & Geis's book. It might be old but it's the best. Have things changed anyway ? For those who want their resume to pass screening: it is clear that showing some knowledge of applications is necessary. If you do not have any, get it. You can read textbooks (see hereafter), you can make connections with labs that do some applied modeling, it will allow you to see if you like that anyway. If you want to know about hot topics, read Science and Nature, Academics and industrials work on the same problems, when you are talking about biological sciences, an example is AIDS (see the last Science issue !) They both would publish in these journals the hot stuff. Also I hate to say that, nothing draws more attention than the lab/university you come from, it is not really different than academia. If your resume shows applications, and your lab is known for techniques such as conformational search, molecular mechanics/dynamics, Poisson-Boltzmann Electrostatics, Free Energy calculations, Homology modeling, QSAR, Distance Geometry, molecular graphics, etc..., there is a good chance you will be perceived as a good candidate. For people with QM background, it really depends more on the specific job offering. Certainly if your work involved essentially diatomic/triatomic molecules, chances are very small. DFT techniques are certainly becoming more and more important though. What I would suggest to people seriously interested in "switching" from a theoretical field to applications in the pharmaceutical industry, is to actually get a post-doc experience in a lab with some recognition in the fields mentioned. I prefer not to mention any, but these are known. For those interested in method development, there are few pharmaceutical companies that will look for that, but the largest ones, or some special cases that have started on the basis of "rational drug design". On the other hand, molecular modeling software companies are normally the most likely to be interested, provided that you can and want to code (preferably in C). As far as I know, only Biosym performs basic research and basic method development among them though, as these relate to applications, usually in collaboration with pharmaceutical companies. Ben Burke from Agouron Pharmaceuticals (burke@tbone.agouron.com ) suggested "that graduate students take Medicinal Chemistry curricula and complement that with additional courses (ugh) in a strong chemistry department (for areas like QM). This should help the student have a reasonable chemistry background, know what it takes to produce (in theory) a successful drug (eg ADME), and have practice modelling ligands and/or protein systems (in the "lab"). The additional course work helps them understand reaction surfaces, transiton states, and geometry-related questions. They would even be on their way to generating necessary parameters, etc. Of course, the quality of the student will depend highly on the mentor as well as the student. The ability to succinctly ask the right question is one that comes from a combination of training (your advisor's at first) and experience generated while working on your project. PS Some statistics training is useful for QSAR work." Gene Fleishman from Cray suggested " If your question really is: What training do I get in computational chemistry to maximize my employmentin the pharmaceutical arena? Then call up some pharmaceutical companies andask them about the experience and training of their computational chemistry staff and while your at it ask about summer internships and work/study programs". Frederick Ignatz-Hoover from Monsanto suggested to "Try J. Chem. Ed., Vol 69, No 7, p. 533 (1992) for a start....", an article on a Molecular modeling teaching course at the Universite de Montreal by Hermann Dugas, which should give a good overview of common molecular modeling techniques of use in industry AND gives basic textbook references. Finally, I would add that apart from the leading companies, who are heavily committed to molecular modeling and have very talented modelers, there is a significant number of companies that did not invest fully in getting BOTH molecular modeling software and molecular modelers. The efficiency of modeling is not due to the software alone. But how many trained molecular modelers are they ? When modeling is central and reliable in a company, is that due to the software, the science behind, or the scientist-user behind ? and to what extent ? I'll let you think about it. For those who asked textbook references (more applications oriented), in addition to the those cited in the J. Chem Ed., 69(7), 1992 paper, I would add a few. It is not an exhaustive list but just a selection. There are plenty of good reviews out there as well (... maybe another time or somebody on the net could volunteer). Specially as one look in the future, the knowledge of protein structure and function is necessary. Is it necessary to say that biology is becoming a branch of chemistry. So shouldn't "computational chemistry" include "computational biology" ? - "Reviews in Computational Chemistry" Lipkowitz & Boyd, eds. Vol. I,II,III VCH publishers: New York - "Computer-Aided Drug Design, Methods and Applications", Thomas J Perun and C.L. Propst, Eds, Marcel Dekker, New York, 1989 - " A Handbook of Computational Chemistry" - a practical guide to chemical structures and energy calculations, Tim Clark, John Wiley and Sons, New York, 1985. - "The structure and action of PROTEINS" Dickerson and Geis, W.A. Benjamin, Inc., 1969 - "Protein Engineering" D.L. Oxender and C. F. Fox Alan R. Liss, Inc, 1987 - "Proteins, A theoretical perspective of dynamics, Structure, and thermodynamics" Charles L. Brooks III, Martin Karplus, B. Montgomery Pettitt, John Wiley, 1988. - "Dynamics of proteins and nucleic acids" by J. Andrew McCammon and Stephen C. Harvey, Cambridge University Press, 1987. - "Prediction of protein structure and the principles of protein conformation" G. D. Fasman Ed. Plenum Press, New York, 1989 - "Density Functional Methods in Chemistry" J. K. Labanowski, J. W. Andzelm, Springer-Verlag (New York : 1991) - "Ab Initio Molecular Orbital Theory", W. J. Hehre, L. Radom, P. v. R. Schleyer, J. A. Pople, Wiley-Interscience Publication (New York : 1986) I hope this helps. Philippe Youkharibache. youkha@biosym.com PS: I'll ge gone until July 1. So if you send me mail I will not be able to respond before then, and even then I'll probably have to trim out hundreds of messages. That's the price to pay with E-mail ! A few original e-mail messages follow: ******************************************************* >From Galen Gawboy (gawboy@sodium.mps.ohio-state.edu) > Philippe, > > >Question for everyone? Is there really a shortage of computational > >chemists/chemists in general? > >>And my answer is : there is an incredible shortage of computational >>chemists in the chemical industry, all over the world. In fact, it is >>much worse in Europe and Japan. (So it is not only a green card issue) > > As someone who is nearing completion of my PhD, and facing a grim >job market for post docs I find this heartening. I for one have absolutely >no idea what industry would look for in a computational chemist. > > I looked in Chemical Engineering News recently and saw only 3 openings >for computational chemists. I have also heard from a head hunter that >personnell department types tend to throw out the theory resume's on the >grounds that they don't see the need to hire someone with "esoteric" skills >with the hope that that person would fit in to the "results oriented" R&D >that their company does. >So if there is huge shortage >of computational chemists where are the openings? How do we sidestep >the personnell "professionals"? What does industry look for? > > Some practical advice of how to get past the resume' screeners and get word >to people who value the skills of computational chemists would be a valued >contribution to the list. For that matter identifying the types of companies >that value computational chemists would be of much use. I tried to have >an informal chat with a representative from a petroleum company to try to >guage what type of general studies their company would be interested in >to try to steer my research direction towards a more real world emphasis. >It turned out that her husband was their group computational chemist. He >has had no training in the field what so ever and seemed to winging it,( I >kept this opinion to myself), surprisingly she was quite hostile towards me. > > Perhaps it is hard for computational chemists to break into industry because >attitudes like this permeate other companies. In other words they hire their >computational chemists from the ranks of their bench chemists and expect them >to pick some black box package and crunch out numbers. I have nothing against >black box packages, but I suspect that they would get more bang for their buck >if they went outside their company and hired somebody who knows how the black >box works. > >Gee, this was longer than I thought it would be. Sorry. > >-galen > ******************************************************* >From markb@orl.mmc.com From: burke@tbone.agouron.com (Ben Burke) From: "Frederick F. Ignatz-Hoover"