From tom@cassandra.chem.washington.edu Sat Mar 13 18:33:09 1993 Date: Sun, 14 Mar 1993 02:33:09 -0800 (PST) From: Thomas Nhan Subject: PDB source file... To: Chemistry Mailing List Message-Id: I'm in the process of writing an application that allow us to view PDB file on our SGI machines. Does anyone know where I can find the source that reads in the PDB files? I've came across the documentation, but finding the source files that read in PDB files will make my life much easier :) Thanks in advance for your help! -- Thomas Nhan | "Genius may have its limitations, but mnhan@u.washington.edu | stupidity is not this handicapped." PGP 2.2 key available | - Elbert Hubbard From doelz@comp.bioz.unibas.ch Sun Mar 14 12:06:13 1993 Message-Id: <199303141006.AA00673@oscsunb.ccl.net> Date: Sun, 14 Mar 93 11:06:13 +0100 From: doelz@comp.bioz.unibas.ch (Reinhard Doelz) To: chemistry@ccl.net Subject: Call for Contributions: Molecular Graphics Announcement of the 2nd Circular and Call for Papers /--\ ================================ \--/ || ... Molecular Graphics Society \ / \ : : \ / \. ... : 12th Annual Conference | . . : | . ...: ================================ / \ / : : / \ / :..: || Interlaken, June 8-11, 1993 /--\ SWITZERLAND \--/ Molecular Graphics and the Design of Bioactive Compounds Theme of the Conference The 12th Annual Conference of the Molecular Graphics Society will focus on the Methods and Results in the Computer Aided Design of Bioactive Molecules, with emphasis on the crucial role which Molecular Graphics plays in the visualization and interpretation of data in this field. The Conference sessions will be centered around the following four general topics 1) Molecular Design (methodology and applications, two sessions) 2) Structure-Guided Design: (a) based on X-ray crystallography (b) based on NMR spectroscopy 3) Databases: Use and Applications 4) Structure-Property Correlations Scientific Programme _____________________________________________ The topic of the conference, "Molecular Graphics and the Design of Bioactive Compounds" has been divided into six half-day sessions, each covering a particular subtopic which is opened by a main lecture followed by a number of selected oral presentations. A long lunch breack is scheduled to enable participants to visit the scientific poster session and the commercial software/handware exhibition. Monday, June 7 afternoon Registration evening Opening of Conference and Welcome Address H.P. Weber, chairman of the Organizing Committee Opening Lecture: Molecular Graphics and Structure Design K. M|ller, F. Hoffmann-La Roche AG, Basel (CH) Welcome Cocktail Tuesday, June 8 morning Session 1: Molecular Design I Main Lecture: 3D Models of Dopamine, Adrenaline, Serotonine, Acetycholine and Mammalian Opsine Receptors Marcel Hibert, Marion Merell Dow, Strasbourg (F) afternoon Session 2: Structure-Property Correlations Main lecture: Hydrophobic Interactions and CoMFA using HINT D.J. Abraham, Virginia Commonwealth University, Richmond VA (USA) Wednesday, June 9 morning Session 3: Structure-Based Design (using NMR and related methods) Main Lecture: At Last Coupling Constants Again - Structure and Dynamics of Peptides and Proteins by NMR and MD Calculations H. Kessler, Technische Universitdt, M|nchen (D) afternoon Conference Excursion, Poster and Commercial Exhibition (optional) Thursday, June 10 morning Session 4: Structure-Based Design (using X-ray Crystallography) Main Lecture: Protein Structure and Drug Design T.L. Blundell, Birkbeck College, London (UK) afternoon Session 5: Databases Main Lecture: The Information Explosion: How Can we Cope? P. Murray-Rust, Glaxo Research, Greenford (UK) evening Conference Dinner Friday, June 11 morning Session 6: Molecular Design II Main Lecture: CAVEAT, TRIAD, and ILIAD P. Bartlett, University of California, Berkeley, CA (USA) Closing Lecture: Computer Simulations of Biomolecules: Improvements in Methodology for Searching Conformational Space and Prediction of Molecular Properties W. van Gunsteren, ETH, Z|rich (CH) afternoon Closing of Conference and Farewell Drink A commercial exhibition of relevant hardware and software will be organized and accessible during the Conference. Scientific Advisory Board Organizing Committee Hoeltje H.D., Berlin Breckenridge R., Basel Karplus M., Harvard Doelz R., Basel Mueller K., Basel Karfunkel H., Basel Richards G., Oxford Van de Waterbeemd H., Basel Silvestre J., Lyon Weber H.P., Basel (Chairman) Thornton J., London Tollenaere J., Beerse Weber J., Geneva Attendance and Call for Contributions Please indicate your interest in attending the Conference and to present a paper by requesting the application form from the conference secretariat 12th Annual Conference of the MGS P.O. Box 6, Clarastrasse 57 CH-4005 Basel, Switzerland Tel. ++41-61-691 51 11 Fax. ++41-61 691 81 89 (Deadline for Abstracts: March 31, 1993) Location and Date of the Conference Interlaken is located in the beautiful alpine Bernese Oberland, nested between the two lovely lakes of Thun and Brienz in breathtaking proximity of the majestic Jungfrau massive. The well equipped Congress Center of Interlaken will be the place of the Conference activities. The Conference activities will start in the evening of Monday, June 7, 1993, and proceed until Friday noon, June 11, 1993. Language The official language of the Conference will be English. No facilities for translation or interpretation will be available. Accommodation and Social Events Interlaken, a traditional alpine health resort, offers a choice of accommodation at various tariffs. A half-day Excursion and a Conference dinner are planned, as well as a programme for accompanying persons. Official carrier: Swissair You are recommended to check with Swissair or your local travel agent for the best flight arrangements for this period. Swissair offices round the world will assist you with your travel arrangements. Further Information If you did not receive the second circular, which was sent out early February, 1993, please contact the conference secretariat at the address given below. The second circular includes registration, abstract and hotel reservation form. 12th Annual Conference of the MGS P.O. Box 6, Clarastrasse 57 CH-4005 Basel, Switzerland Tel. ++41-61-691 51 11, Fax. ++41-61 691 81 89 +----------------------------------+-------------------------------------+ | Dr. Reinhard Doelz | RFC doelz@urz.unibas.ch | | Biocomputing | DECNET 20579::48130::doelz | |Biozentrum der Universitaet | X25 022846211142036::doelz | | Klingelbergstrasse 70 | FAX x41 61 261- 6760 or 267- 2078 | CH 4056 Basel | TEL x41 61 267- 2076 or 2247 | +------------- bioftp.unibas.ch is the SWISS EMBnet node ----------------+ ----------------------------------------- From m10!frisch@uunet.UU.NET Sun Mar 14 10:26:38 1993 Message-Id: <9303142141.AA05900@relay1.UU.NET> Date: Sun, 14 Mar 93 15:26:38 EST From: m10!frisch@uunet.UU.NET (Michael Frisch) Subject: Isotopic substitution on g92 To: chemistry@ccl.net Louis Grace asked, As a reference for some benzenoid compounds I'm studying, I'm doing ab initio calculations (using Gaussian92) on benzene and benzene-d6. I've successfully run the benzene molecule, but when I try a frequency calculation on benzene-d6, g92 doesn't seem to like my input file. It keeps giving me "Expecting floating point, found end of line." Does anyone know the exact format of an input file for a frequency calculation with isotopic substitution (using the keyword, freq=readisotopes) under g92? Also, my benzene molecule was built with a dummy variable in the center, and it is the first line in the z-matrix. Do I just ignore this when putting in my isotopic masses? The input for Freq=ReadIso consists of one line with the parameters for the thermochemistry -- Temperature (Kelvin), Pressure (Atm), and (starting with Gaussian 92 Revision D) optionally the scale factor to apply to the frequencies. Zero temperature or pressure defaults to the standard values. A zero scale factor defaults to the recommended value of 1/1.12. Omitting the scale factor or setting it to 1 leaves the frequencies unchanged. The masses are entered as integers, with one for each atom on a line by itself. Only real (i.e., non-dummy) atoms are specified, as the integer values customarily used to refer to the isotopes. The program then uses the corresponding exact value -- for eaxmple, requesting a mass of 18 for Oxygen causes the value 17.9915939 to be used. Here's a complete input file for water frequencies with the isotopes O18, D, and T. A temperature of 100 Kelvin, 2 atmospheres pressure, and the standard scale factor are also requested: # rhf/sto-3g freq=readiso water 0 1 O H,1,R2 H,1,R2,2,A3 R2=0.98940918 A3=100.02691259 100.0 2.0 0.0 18 2 3 Note that the values printed for the frequencies are the unscaled ones, even if the thermochemistry uses scaling.y In versions of Gaussian 92 prior to Revision D, the scale factor cannot (and should not) be specified. Mike Frisch frisch@lorentzian.com ------- From CUNDARIT@MSUVX1.MEMST.EDU Sun Mar 14 13:20:33 1993 Date: 14 Mar 1993 19:20:33 -0600 (CST) From: CUNDARIT@memstvx1.memst.edu Subject: LDF followup To: chemistry@ccl.net Message-Id: <01GVT5IY0R8Y9UM7FR@MSUVX1.MEMST.EDU> Fellow Net-minders, A week or so ago, I posted some question about density functional calcs. that we are playing with. The original posting follows this intro. In summary...1) LDF methods are size-consistent. 2) The frozen core approximation should have minimal effect on geometry and energetics (we did the expt. and for these cases there is no diff). 3) For accurate energetics the biggest concern revolves around the use of non-local/gradient corrections. Tom Cundari, Asst. Prof. Dept. of Chemistry Memphis State University Memphis, TN 38152 phone:901-678-2629 fax:901-678-3447 e-mail:cundarit@memstvx1.memst.edu *************************** original mail ************************* Hi, Our group has just started experimenting with LDF in a small way for transition metal calcs. We are trying to calculate the binding energy of H2 to Cr(CO)5, in conjunction with some work by Ted Burkey in our Dept. who has made the first exptl. determination of this quantity. Given our lack of experience with LDF we want to be sure we are not making any serious mistakes. We are using DMol at the San Diego Supercomputer Center. First question: Are LDF methods size-consistent? Second question: Does anybody have any feeling for the effect of freezing the core density in such calcs. on properties such as energetics and geometry? Third question: I have just calcd. a binding energy of -30 kcal/mol from E(Cr(CO)5(H2)) - E(Cr(CO)5) -E(H2) using the equivalent of a double-zeta- valence-plus-polarization basis set, and a FINE mesh size. Should I believe this or is it worth the expense to go to a larger basis set, finer mesh, unfrozen core? All suggestions are welcome. From: IN%"JSMCM@royal.crc.uno.edu" "JORGE M. SEMINARIO" 6-MAR-1993 17:15:14.87 > First question: Are LDF methods size-consistent? Yes, but it does not guaranteed good energetics. The precise answer to your question is in the chapter: Size-Consistency, Self-interaction Correction, and Derivative Discontinuity in DFT by John P. Perdew, Advances in Quantum Chemistry (vol.21) 1990. The whole volume is dedicated to DFT with very interesting papers. > Second question: Does anybody have any feeling for the effect of >freezing the core density in such calcs. on properties such as energetics >and geometry? I can not say anything about frozen core. But the energetics is already very poor using full core local approximations. We have found out that the errors are so large for molecules containing up to 4 four first row atoms that definitely no energetics can be obtained from the local approximation. However, we have a practical way to correct the energies so good energetics can be obtained. I can send you a preprint (in press in CPL) if you are interested. The situation right now is that very sophisticated and precise non-local functionals are very common. They produce energetics as good as the best ab initio calculations. The situation might be different for transition metals, which we are investigating right now using the Perdew-Wang functionals. Regarding the geometries, LDF seems to yield very good ones; anyhow, the non-local ones are always better. > Third question: I have just calcd. a binding energy of -30 kcal/mol >from E(Cr(CO)5(H2)) - E(Cr(CO)5) -E(H2) using the equivalent of a double-zeta- >valence-plus-polarization basis set, and a FINE mesh size. Should I believe >this or is it worth the expense to go to a larger basis set, finer mesh, >unfrozen core? My suggestion, would be to have a first a good non-local or/and ab-initio as a reference if experimental data is not available. Best regards Jorge Seminario PS. I would appreciate to have summary of the answers. From: IN%"mckelvey@Kodak.COM" 4-MAR-1993 16:51:23.48 If you are not already doing it, I would consider non-local/gradient corrections. This should be available about now. You might look up recent work by Tom Ziegler (Calgary), in JACS and elsewhere on this subject. He has shown the non-local part to be of extreme importance. John McKelvey Kodak Res Labs 716-477-3335 From: IN%"george@archimedes.cray.com" 4-MAR-1993 14:24:45.20 Prof. Cundari, Here is my best guess at the answers to your DFT questions: 1) DFT should be size-consistent in the CI sense. However, the electron gas approximation may not be an equally good approximation for a molecule and its fragments. This can introduce errors if local DFT is used. 2) If you freeze the core on the Cr, this should have minimal impact. 3) Including non-local corrections is much more important than improving the grid or the basis set. This is espcially true for computing energetics. THe local results are pretty good for structures, but accurate energies require non-local corrections. These apparently will be available in DMol 2.3. Non-local corrections are also available in DGauss (from Cray), deMon (from Dennis Salahub), and AMol (from E.J. Baerends). George Fitzgerald Cray Research From: MSUVX1::WINS%"salahub@ERE.UMontreal.CA" 4-MAR-1993 09:38:08.28 Density functional methods are size consistent. I think the grid and basis setr are likely OK. The key question is whether you are using the nonlocal gradient correctd functional (available in only the latest version of DMOL ). The LDA usually overestimates binding energies seriously. You could ask the folks at BIOSYM if you need the nonlocal version of DMOL. It would also be possible to use deMon for this kind of problem. This is also a BIOSYM product. Please don't hesitate to ask if you think I could help . Sincerely yours, Dennis Salahub From Louis.Grace@um.cc.umich.edu Sun Mar 14 16:53:26 1993 Date: Sun, 14 Mar 93 21:53:26 EST From: Louis.Grace@um.cc.umich.edu To: chemistry@ccl.net Message-Id: <20913720@um.cc.umich.edu> Subject: More g92 isotope problems Hello, network! I would like to thank everyone who responded to my question about format of the input file to do isotopic substitution for a frequency calculation under Gaussian92(TM). My problem arose from the fact that the g92 brochure's example merely states that in g92 one does not need the "NO" line. It does not mention, though, that one needs the temperature and pressure! When I put these in, it seemed to read the file o.k. Thanks! Now I have another problem. While g92 seems to be getting all the atom coordinates, and it tells me it has read the isotopic masses, the force constants are all ZERO! I have been using the following command line: (sorry, lines:) "%chk=benzene.chk" "#hf/3-21g test freq=(readfc,readisotopes) guess=read geom=check" (The quotes are because I'm on an IBM mainframe with a STUPID operating system that tries to interpret the % sign as a command!) Anyway, the output seems to contain the standard orientation and reduced masses, but, as I said, the force constants, energies and frequencies are all zero. What could be the problem? The checkpoint file was generated in a run for which the command line was just: #hf/3-21g opt test Did I need anything else? Could the checkpoint file be screwed up? Is there possibly another keyword I'm missing in the isotope run? Thanks in advance to anyone who can help. Louis Grace Chemistry Department The University of Michigan Ann Arbor, Michigan 48104