From owner-chemistry@ccl.net Mon Dec 4 05:55:34 1995 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.6.10/950822.1) id FAA08528; Mon, 4 Dec 1995 05:53:28 -0500 Received: from uibk.ac.at for Michael.Probst@uibk.ac.at by bedrock.ccl.net (8.7.1/950822.1) id FAA04471; Mon, 4 Dec 1995 05:53:18 -0500 (EST) Received: from dm.uibk.ac.at (dme.uibk.ac.at) by uibk.ac.at with SMTP id AA26677 (5.65c/IDA-1.4.4 for ); Mon, 4 Dec 1995 11:53:09 +0100 Received: from mp3-c724 by dm.uibk.ac.at (950911.SGI.8.6.12.PATCH825/CFIBK-2f.IRIX) id LAA24253; Mon, 4 Dec 1995 11:53:09 +0100 Message-Id: <199512041053.LAA24253@dm.uibk.ac.at> Comments: Authenticated sender is From: "Michael Probst" To: CHEMISTRY@ccl.net Date: Mon, 4 Dec 1995 11:53:10 +0000 Subject: Summary: Gaussian 94 opt+freq Reply-To: michael.probst@uibk.ac.at Cc: c72417@dm.uibk.ac.at Priority: normal X-Mailer: Pegasus Mail for Windows (v2.01) Hi ! A while ago I posted a question regarding the fact that a Gaussian 94 job which contained optimization and frequency calculation in the same deck did not give the same energy at the end of the optimization step and in the frequency step. It turned out that there was a bug in Gaussian 94 which used the aug-cc..... basis set as requested in the optimization step but 'forgot' the 'aug' and used the cc-.... basis set in the frequency step. I was told by Gaussian that this bug was known and has been corrected in new g94 releases (I used SGI-G94rev.B.3). I want to thank Christopher Cramer who noticed this bug before. I got lots of mail from people who did not know that specifying both freq and opt on the control line is possible in g94. (In g92 it would have been an error) Thanks to all. Michael michael.probst@uibk.ac.at --------------------------------------------------------------------------- Michael Probst Internet: michael.probst@uibk.ac.at Department of Inorganic Chemistry Fax: ++43(512)507 2934 Innsbruck University Tel: ++43(512)507 5153 Innrain 52a Austria From C867BUC%SEMOVM.BITNET@phem3.acs.ohio-state.edu Mon Dec 4 10:40:38 1995 Received: from phem3 for C867BUC%SEMOVM.BITNET@phem3.acs.ohio-state.edu by www.ccl.net (8.6.10/950822.1) id KAA12895; Mon, 4 Dec 1995 10:38:26 -0500 Received: from SEMOVM (C867BUC@SEMOVM) by phem3.acs.ohio-state.edu (PMDF V4.2-13 #5888) id <01HYEN960P0099E3YC@phem3.acs.ohio-state.edu>; Mon, 4 Dec 1995 10:38:17 EST Date: Mon, 04 Dec 1995 09:35:41 -0600 (CST) From: "Anthony J. Duben" Subject: VAX/MicroVAX hardware and software To: chemistry@www.ccl.net Message-id: <04DEC95.10362618.0077.MUSIC@SEMOVM> X-Mailer: MUSIC/SP V4.1.0 Content-transfer-encoding: 7BIT An industrial donor offered to donate the following items to my institution, and I would like to have the opinion of those on the list whether we would benefit. The donor wishes to donate the followng computers: VAX 4000-200 MicroVAX 3500 MicroVAX 3300 It seems that we will not be getting the system software and compilers. We would need to get them from DEC or another vendor. I am aware that they are not state of the art, but they might have some use for compute bound calculations -- set up the calculation and let them crank for a couple of weeks. I would appreciate opinions of those on the list. Is this a deal or are they boat anchors? Please e-mail direct instead of to the list to avoid clutter. ********************************************************************** Anthony J. Duben Southeast Missouri State University Associate Dean, College of Science and Technology Rhodes Bldg. 102C, Mail Stop 6000 1 University Plaza Cape Girardeau MO 63701-4799 voice: (314) 651-2194 FAX: (314) 651-2223 e-mail: C867BUC@SEMOVM.SEMO.EDU ********************************************************************** From rosas@irisdav.chem.vt.edu Mon Dec 4 12:10:39 1995 Received: from irisdav.chem.vt.edu for rosas@irisdav.chem.vt.edu by www.ccl.net (8.6.10/950822.1) id MAA14797; Mon, 4 Dec 1995 12:00:42 -0500 Received: by irisdav.chem.vt.edu (940816.SGI.8.6.9/940406.SGI) for chemistry@www.ccl.net id LAA10143; Mon, 4 Dec 1995 11:45:35 -0500 From: "Victor M. Rosas Garcia" Message-Id: <9512041145.ZM10141@irisdav.chem.vt.edu> Date: Mon, 4 Dec 1995 11:45:31 -0500 X-Mailer: Z-Mail (3.2.0 26oct94 MediaMail) To: chemistry@www.ccl.net Subject: SUMMARY-Global minimum philosophical stuff Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Hi there, Well, it took me some time to assimilate all the answers I got to my original posting (see below for the original question). I received a good number of answers that I summarize and integrate below. Thanx to: Georg Schreckenbach Mike Gilson Gregory V. Nikiforovich Rene Fournier Jan H. Jensen Robert W. Zoellner Nathalie Meurice Matt (/G=Matthew/S=Harbowy/OU=LIPTONUS-EC02/O=TMUS.TJL/@LANGATE.gb.sprint.com and Eugene Leitl for all their input. My original posting was as follows: ******************************************************************************** Hi everybody, I'm just getting into the free energy calculation business, so I've been reading the most basic references (by van Gunsteren, Jorgensen, Mitchell and people like them). The first thing I noticed is that it is extremely important to make sure that one has a global minimum, together with all the conformers within certain energy; and it is very obvious that it is not easy to make sure that a conformer is "the" global minimum. I wonder if it is theoretically IMPOSSIBLE to prove that a given conformation is the global minimum. In my mind, is the same problem found when we want to say that we have found *all* the conformers within a certain energy range. Both (I think) boil down to proving a negative proposition (something that we *logically* cannot do), e.g. For the global minimum: "There are no other conformers with a lower energy than this one." For a distribution: "There are no other conformers that should be included in the population." Of course, it is always possible to *disprove* or *falsify* a negative proposition by showing a counterexample, but being unable to find a counterexample doesn't prove the truth of a negative proposition. OTOH, both of the propositions above can be cast in an affirmative form if we restrict them to a particular case, e.g.: For the global minimum, we could easily prove or disprove: "This particular conformer is higher in energy than the one we call 'global minimum'." By the same token, for a distribution: "This particular conformer has an energy higher than our cutoff limit of interest, (so we don't have to include it in the distribution)." So, I guess the point I'm trying to make is that we will *never ever* be sure of having "the" global minimum or a complete population because it is logically impossible to prove it. By running several conformational searches with different starting conditions we are probably doing the best we can. I'd like to know your opinions on this matter. PS Sorry if this stuff is too obvious, it isn't to me. Victor ******************************************************************************** I think I can divide the topics addressed in the answers in three categories: a) problems in making sure that we found a global minimum b) approaches to make sure that we found a global minimum c) applications a) problems in making sure that we found a global minimum In general, people agree that it is very difficult to assess if we have a global minimum from the mathematical point of view. Examples show that certain functions can be very pathological and it can be very difficult to find a minimum. Fortunately, in chemistry we are restricted to functions of a particular kind, and those are usually well behaved. Even in that case, it is not so much the form of the functions that makes the problem difficult, but the huge number of minima in molecules like proteins is what kills us. Basically it is a war of attrition of the proteins against our algorithms, and the proteins have always won, so far :). A point was made that if we knew the shape for the whole energy hypersurface it *would* be possible to say which one is the global minimum. I guess that answers my question that *in principle* (logically), it is possible to say that a global minimum has been found. Now I'm thinking about the same question as applied to statistical methods for the search, precluding any knowledge of the energy hypersurface. b) approaches to make sure that we found a global minimum There were as many different approaches as answers, I'd say. The methods mentioned were: -Run several statistical searches with different starting points, counting repetitions of the lowest energy conformer until some criterion for repetition is satisfied. The problems here are how to quantify the level of confidence for a given conformer to be the global minimum, and how different conformers have to be to be considered "different". - Use of genetic algorithms. - Use of Explicit Statespace Trasversal Modeling which is: >An attempt to model the system by representing it by a >vector of numbers to be parallelly (there's a trick to it) changed >by a transition table. The local Hamiltonian is computed each time >anew from elementary contributions. This approach would use Cellular Automata-like encoding. - Also Genetic programming calibration (this is very related to genetic algorithms) An interesing point was that energy minimization is just a way to find a 3D arrangement. A comment to this is that the set of 3D arrangements is a superset of the set of conformers and it includes isomers of the molecule of interest. Usually one tries to avoid isomerizations during the search. c) Applications This got to be interesing, in the particular case of proteins, some people questioned the validity of the assumption that proteins are at the global minimum. After checking a couple of references (mostly Anfinsen's seminal paper in Science and Ken Dill's review in Biochemistry) I sort of confirmed that there is little evidence for that. So the search for the global minimum of a protein *could* be a waste of time. Also, an important point was raised, in that it is more important to consider the applicability of the geometries to experimental problems, than to be concerned about the ranking in energy. I think that's it. I hope summary was clear and that I did not misunderstood the answers I got. See ya' Victor -- ----------------------------------------------------------------------- Victor M. Rosas Garcia * "How can we contrive to be rosas@irisdav.chem.vt.edu * at once astonished at the Virginia Tech doesn't necessarily share * world and yet at home in it?" the opinions you just read. * G. K. Chesterton ------------------------------------------------------------------------- From yuan@nka1.med.uc.edu Mon Dec 4 12:40:40 1995 Received: from nka1.med.uc.edu for yuan@nka1.med.uc.edu by www.ccl.net (8.6.10/950822.1) id MAA15497; Mon, 4 Dec 1995 12:40:04 -0500 Received: by nka1.med.uc.edu (920330.SGI/920502.SGI.AUTO) for chemistry@www.ccl.net id AA07943; Mon, 4 Dec 95 12:46:26 -0500 Date: Mon, 4 Dec 1995 12:46:18 -0500 (EST) From: Jie Yuan To: Computational Chemistry List Subject: wanted: 8th Intl. Conf. on Na,K-ATPase info Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII I sent email to the organizer but it was returned for the domain name was not recognized. Would someone send me some information about this conference? The deadlines for submission is the most important at this time. Again, the name is: The 8th International Conference on the Na,K-ATPase, in Mar del Plata, Argentina, in Aug '96. Many thanks! Jie -- Jie Yuan, PhD - Pharmacology & Cell Biophysics - U. Cincinnati -- -- Phone: 513-558-2352 -- Fax: x-1169 -- Email: Jie.Yuan@UC.Edu -- -- P.O. Box 670575, 231 Bethesda Ave., Cincinnati, OH 45267-0575 -- -- http://www.uc.edu/~yuanj, http://www.pharm.med.uc.edu -- From owner-chemistry@ccl.net Mon Dec 4 13:10:40 1995 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.6.10/950822.1) id NAA15812; Mon, 4 Dec 1995 13:10:08 -0500 Received: from chx400.switch.ch for rytz@solar.iac.unibe.ch by bedrock.ccl.net (8.7.1/950822.1) id NAA11391; Mon, 4 Dec 1995 13:10:04 -0500 (EST) Received: from arwen.unibe.ch by chx400.switch.ch with SMTP (PP); Mon, 4 Dec 1995 19:09:37 +0100 From: rytz@solar.iac.unibe.ch (Ruedi Rytz) Message-Id: <9512041809.AA22764@solar.iac.unibe.ch> Subject: New EHMO and Oscillator Strength Package To: chemistry@ccl.net (Computational Chemistry List) Date: Mon, 4 Dec 1995 19:09:23 +0100 (NFT) X-Mailer: ELM [version 2.4 PL20] Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit We can offer you the DOS/Windows3.1 binaries of ICON-EDiT an Extended-Hueckel and Oscillator Strength Calculation Package by Gion Calzaferri, Ruedi Rytz, and Martin Braendle Institute for Inorganic and Physical Chemistry Freiestrasse 3, 3012 Berne, Switzerland Copyright by the Authors -------------------------------------------------------------------------------- THE FULL VERSION FOR IBM RS/6000 AS WELL AS FOR DOS/WINDOWS ALONG WITH A DETAILED MANUAL WILL BE MADE AVAILABLE BY MARCH 1996. IT WILL BE DISTRIBUTED ON DISKETTES TO A PRICE OF US$ 60 TO COVER OUR EXPENSES. YOU CAN ODER IT BY SENDING A FAX TO PROF. DR. G. CALZAFERRI (FAX: +41 31 631 39 94). DO NOT FORGET TO INCLUDE YOUR SHIPPING ADDRESS. -------------------------------------------------------------------------------- ICON-EDiT is a FORTRAN program package that performs extended-Hueckel and oscillator strength calculations on molecules. s,p,d orbitals, a two-body repulsive energy term, different Wolfsberg-Helmholz formulas, charge iteration procedures, geometry variation and an FMO option are included. ICON-EDiT consists of three parts, INPUTC, ICONC, and EDiT. The INPUTC part of the package is a self-explanatory input program. It reads the necessary parameters such as the valence orbital ionization energies (VOIEs), the Slater-type orbitals (STOs) as well as the charge-iteration parameters from the ASCII data files ATOMDEF.DAT, VOI.DAT, and FOI.DAT in order to generate the file SOMENAME.KAR (Cartesian coordinates) or SOMENAME.GEN (internal coordinates with geometry variation). As ICONC is designed to work with Cartesian coordinates, the previously generated internal coordinate file SOMENAME.GEN can be converted to a Cartesian file named TEMP.KAR (cf. INPUTC Option #3). The parameter files may be extended and/or adapted to meet the requirements of the user. ICONC performs extended-Hueckel calculations on files created by INPUTC. EDiT allows to calculate oscillator strengths of electronic dipole-induced transitions (EDiTs) based on Slater-type extended-Hueckel wave functions. You can get ICON-EDiT from the anonymous ftp server "iacrs1.unibe.ch" (130.92.11.3). You will find the 32-bit binaries (iconedit.zip) along with a few representative examples (examples.zip) in pub/ICON-EDiT. More information is included in the file README.1ST. Enjoy the programs Ruedi Rytz and Gion Calzaferri -- -------------------------------------------------------------------------------- Ruedi Rytz | Welcome my son, welcome to the machine. | Where have you been? Institute for Inorganic and | It's alright we know where you've been. Physical Chemistry | You've been in the pipeline... | University of Berne | Freiestr. 3, CH-3012 Bern | Pink Floyd E-Mail rytz@solar.iac.unibe.ch | Wish you were here Tel. +41 31 631 42 25 | Fax +41 31 631 39 94 | -------------------------------------------------------------------------------- From 94970459@vax1.dcu.ie Mon Dec 4 16:40:42 1995 Received: from VAX1.DCU.IE for 94970459@vax1.dcu.ie by www.ccl.net (8.6.10/950822.1) id QAA19859; Mon, 4 Dec 1995 16:24:51 -0500 Received: from vax1.dcu.ie by vax1.dcu.ie (PMDF V4.2-12 #4433) id <01HYF51DS3TS96YBM9@vax1.dcu.ie>; Mon, 4 Dec 1995 19:26:39 GMT Date: Mon, 04 Dec 1995 19:26:39 +0000 (GMT) From: Paddy Kane <94970459@vax1.dcu.ie> Subject: delta G calculations To: chemistry@www.ccl.net Message-id: <01HYF51DT6EQ96YBM9@vax1.dcu.ie> X-Envelope-to: chemistry@www.ccl.net X-VMS-To: IN%"chemistry@www.ccl.net" MIME-version: 1.0 Content-transfer-encoding: 7BIT Hi CCL'ers I am a HyperChem user and I wish to carry out calculate delta G of complexationfor complexes formed between calixarenes and metal ions. HyperChem cannot handle delta G calculations, so I enquired about the MOPAC software as it can open files in *.zmt format and HyperChem can save files in this format. However Richard Counts of the QCPE informs that MOPAC can handle molecules with at most 80 hydrogens and at most 80 non-hydrogens. However the molecules of interest to me typically have at least 100 of both types of atoms. So, what I would like to know if there is software available which a) can carry out delta G calculations on the complexes of interest to me b) is compatible with MS-DOS 6.2 or WINDOWS 3.1 or WINDOWS 95 c) opens files in a format compatible with HyperChem Obviously, the first two requirements are essential while the last one is desirable. Regards, Paddy. Paddy Kane School of Chemical Sciences Dublin City University Ireland 94970459 e.mail 94970459@vax1.dcu.ie From owner-chemistry@ccl.net Mon Dec 4 17:10:42 1995 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.6.10/950822.1) id RAA20378; Mon, 4 Dec 1995 17:06:08 -0500 Received: from donald.uoregon.edu for campos@OREGON.UOREGON.EDU by bedrock.ccl.net (8.7.1/950822.1) id RAA13717; Mon, 4 Dec 1995 17:06:03 -0500 (EST) Received: from gerardo.uoregon.edu (gerardo.uoregon.edu) by OREGON.UOREGON.EDU (PMDF V5.0-5 #13764) id <01HYEUIKHGFK8Y52JH@OREGON.UOREGON.EDU> for chemistry@ccl.net; Mon, 04 Dec 1995 14:05:55 -0800 (PST) Date: Mon, 04 Dec 1995 14:07:09 -0300 From: campos@OREGON.UOREGON.EDU (Gerardo Soto-Campos) Subject: SUMMARY WASP X-Sender: campos@oregon.uoregon.edu To: chemistry@ccl.net Message-id: <01HYEUIKJOHU8Y52JH@OREGON.UOREGON.EDU> MIME-version: 1.0 X-Mailer: Content-type: text/plain; charset="us-ascii" Content-transfer-encoding: 7BIT A couple of days ago I requested info about the software WASP; and some people asked me to post a summary. I got one answer, which follows: **************************************************************************** ************ Dominic Di Toro wrote to me: Dear Dr. Soto-Campos I have no idea if the WASP you are referring to is the Water Analysis and Simulation Program (WASP) that we authored in early 1980's and is still used today in modified form. It is a water quality simulation program that is general enough to be used for streams, estuaries, lakes and oceans. It computes distributions of dissolved oxygen, and eutrophication related variables: nitrogen forms, phosphorus and algal chlorophyll. Let me know if this is what you are interested in. Dominic M. Di Toro Research Professor Environmental Engineering Department Manhattan College Bronx NY 10471 dditoro@hydroqual.com **************************************************************************** ******** Also from Dominic Di Toro: "Contact the EPA lab in Athens, Georgia. They will supply you with code, manuals etc. Bob Ambrose: 404-546-3546 or Tom Barnwell: 404-546-3175 Also, see the EPA home page and look for the Athens Research lab. I've been told that you can get the code by FTP. Have fun, Dom Dominic M. Di Toro Research Professor Environmental Engineering Department Manhattan College Bronx NY 10471 dditoro@hydroqual.com **************************************************************************** ************** I hope is useful to the people who requested it. Sincerely, Gerardo Soto-Campos. From LSTEFFEN@fair1.fairfield.edu Mon Dec 4 21:10:45 1995 Received: from fair1.fairfield.edu for LSTEFFEN@fair1.fairfield.edu by www.ccl.net (8.6.10/950822.1) id VAA22811; Mon, 4 Dec 1995 21:00:05 -0500 From: Received: from fair1.fairfield.edu by fair1.fairfield.edu (PMDF V4.3-7 #10178) id <01HYF5OZQSOG938HBT@fair1.fairfield.edu>; Mon, 4 Dec 1995 19:33:30 EST Date: Mon, 04 Dec 1995 19:33:30 -0500 (EST) Subject: N cubed processor To: chemistry@www.ccl.net Message-id: <01HYF5OZSEK2938HBT@fair1.fairfield.edu> X-Envelope-to: chemistry@www.ccl.net X-VMS-To: IN%"chemistry@www.ccl.net" MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Content-transfer-encoding: 7BIT We are considering the possibility of acquiring an N^3 parallel processor with an IRIX front end.(32 proc)I would like to hear from people in the computational chemistry community who have such a system and wether or not they have been satisfied with its performance? Is it significantly faster than regular processors? Were you able to install and use G90, 92 or a similar program easily? Sorry for a rather general question, but I really wanted some input >from computational chemists who have actually used this box before getting a possible expensive to maintain and use paperweight!! Please send responses directly to me at: lsteffen@fair1.fairfield.edu or lsteffen@aol.com L. Kraig Steffen Assistant Professor of Chemistry Fairfield University Fairfield Connecticut, 06430 203 254 4000 ex 2254 fax: 203 254 4034