From ep7@dent.okayama-u.ac.jp Fri Oct 25 02:31:56 1996 Received: from deews1.dent.okayama-u.ac.jp for ep7@dent.okayama-u.ac.jp by www.ccl.net (8.8.0/950822.1) id CAA28862; Fri, 25 Oct 1996 02:01:03 -0400 (EDT) From: Received: from [150.46.140.80] (yobou2.dent.okayama-u.ac.jp) by deews1.dent.okayama-u.ac.jp (5.67+1.6W/6.4JAIN-1.1) id AA18187; Fri, 25 Oct 96 15:00:15 JST Message-Id: <9610250600.AA18187@deews1.dent.okayama-u.ac.jp> Date: Fri, 25 Oct 1996 15:08:31 +0900 To: chemistry@www.ccl.net Subject: CCL:SUMMARY:weekness for atomic cahrge using fitting to electrostatic potential methods Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) I send you a summary of answers about ESP-charges.　I would like to thank everybody for many advices. I hope this summary will help everybody. Masao Masamura Date: Sun, 29 Sep 96 13:30:49 +0100 From: Konrad Hinsen To: ep7@dent.okayama-u.ac.jp Subject: Re: CCL:weekness for atomic cahrge using fitting to electrostatic potential methods There was a similar question a few days ago, so I'll just send you the reply I sent then: Date: Wed, 25 Sep 96 20:19:42 +0100 From: Konrad Hinsen To: hutschka@quantix.u-strasbg.fr Cc: chemistry@www.ccl.net In-Reply-To: <9609251527.AA55255@quantix.u-strasbg.fr> (hutschka@quantix.u-strasbg.fr) Subject: CCL:G:Correlation effect on calc atomic Charges Sender: Computational Chemistry List Errors-To: ccl@www.ccl.net Precedence: bulk > I'm calculating atomic charges with Gaussian using Mulliken population > analysis and fitting to electrostatic potential methods. You didn't say *why* you are doing this, so I assume that the goal is obtaining partial charges for an empirical force field. > I've noted some important differences between Mull and ESP derived charges. > More , for ESP derived charges the use of correlated densities (MP2 density >in this case) > gives important differences with the use of HF density. First of all, Mulliken charges are based on a somewhat arbitrary assignment of parts of electron densities to the individual atoms. They may give some indication of polarization etc., but as a source for partial charge information they are not very useful. ESP-derived charges depend on 1) the ab-initio reference potential (which includes basis-set dependence) 2) the choice of evaluation points for the potential 3) the method used for fitting the charges. Dependence 1) is obvious and if you find that your charges depend on the level of calculation (provided that you haven't done something stupid regarding 2) and 3)), then this is a feature of your system that you have to understand and draw conclusions from. The dependence on the evaluation points is equally obvious, but there is no single choice that everyone agrees on (and to some extent it depends on what you want to do with the fitted charges). Among the strategies that have been proposed are: 1) points on a grid around the molecule (not recommendable due to the dependence on arbitrary grid axes) 2) points on well-defined surfaces around the molecule 3) points chosen at random in a well-defined region around the molecule. Dependence 3) is more problematic, since it shouldn't be there in an ideal world. The proposed methods differ in numerical stability and in the exact quantity they are trying to minimize. Basically, everyone agrees that what we want is a least-squares fit. Such fits are in general known to be problematic, because the solution is often underdetermined, and this also occurs for charge fitting. It is therefore *not* a good idea to simply solve the normal equations for the least-squares, but unfortunately that is what most people are doing. A much safer alternative is singular-value decomposition, as described in most books on matrix computations, or even in the second edition of Numerical Recipes. Beyond the problem of numerical stability of the fit, one popular method (known as RESP) proposes to put a constraint on the absolute value of the charges, based on the observation that ESP often gives charges that seem to large. I don't know in how far this is a result of numerical instabilities (RESP does not use singular value decomposition) or of some real physical effect; this ought to be investigated. Literature: C.I. Bayly, P. Cieplak, W.D. Cornell and P.A. Kollman J. Phys. Chem. 97, 10269 (1993) (This paper describes RESP.) K. Hinsen and B. Roux J. Comp. Chem., in print (contact me for a preprint) (This paper describes a specific potential function for proton transfer simulations, but contains an extensive appendix that describes a charge-fitting strategy based on singular-value decomposition.) Both papers contain references to older methods. There is another paper that has appeared earlier this year in J. Comp. Chem. and which deals explicitly with an SVD-based fitting method, but I haven't seen it yet. ------------------------------------------------------------------------------- Konrad Hinsen | E-Mail: hinsen@ibs.ibs.fr Laboratoire de Dynamique Moleculaire | Tel.: +33-76.88.99.28 Institut de Biologie Structurale | Fax: +33-76.88.54.94 41, av. des Martyrs | Deutsch/Esperanto/English/ 38027 Grenoble Cedex 1, France | Nederlands/Francais ------------------------------------------------------------------------------- I thank you for your replay. A several years ago, I found that the CHELP cannot reproduce the atomic charge for HCOO-(H2O)n (n = 0,1,2,3,4,5,6). I think CHELP cannot reproduce the atomic charge for clusters. I will publish these results. Thank you. ============================================================================ ===== Masao Masamura Okayama University Dental School Department of Preventive Dentistry Fax: 81-86-225-3724 e-mail: ep7@dent.okayama-u.ac.jp ============================================================================ ===== Date: Mon, 30 Sep 96 10:28:30 +0100 From: Konrad Hinsen To: ep7@dent.okayama-u.ac.jp Subject: Re: Thanks > A several years ago, I found that the CHELP cannot reproduce the > atomic charge for HCOO-(H2O)n (n = 0,1,2,3,4,5,6). I think CHELP cannot > reproduce the atomic charge for clusters. I will publish these results. What do you mean by "CHELP cannot reproduce the atomic charge"? Which atomic charge? There is no measurable or even well-defined quantity called "atomic charge". In reality, you have well-localized nuclear charges plus strongly delocalized electronic charge distributions. The point of ESP procedures such as CHELP is to reproduce the electrostacic potential of this complicated charge distribution. -- ------------------------------------------------------------------------------- Konrad Hinsen | E-Mail: hinsen@ibs.ibs.fr Laboratoire de Dynamique Moleculaire | Tel.: +33-76.88.99.28 Institut de Biologie Structurale | Fax: +33-76.88.54.94 41, av. des Martyrs | Deutsch/Esperanto/English/ 38027 Grenoble Cedex 1, France | Nederlands/Francais ------------------------------------------------------------------------------- I thank you for your valuable opinion. The CHELP cannot reproduce the charge on H' of H'COO-(H2O)n (n=0,1,2,3,4,.5.6) as follows: When n becomes larger, Natural population analysis predict the charge on the H' becomes more positive (J.Phys.Chem., 97, 3157(1993)). I think the prediction of Natural population analysis is correct for the following reasons: I showed the cause for the elongation of the H-C bond in HCOO- in the gas phase. This elongation results from the contribution of H-...CO2 as resonance structure to the HCOO-. When more water molecules attach HCOO-, more minus charge withdraws from HCOO- to water molecules. Consequently, it is predicted that the resonance structure, H-...CO2, contributes less to HCOO- with n increment. Due to less contribution of the resonance structure H-...CO2 with each n increment, the charge on the H' becomes more positive with n increment. However, the CHELP disagrees with this prediction. That results for CHELP is old. Thus, I will recalculate the charge on the H' with GAUSSIAN 94. Also, I will use CHELPG. ============================================================================ ===== Masao Masamura Okayama University Dental School Department of Preventive Dentistry Fax: 81-86-225-3724 e-mail: ep7@dent.okayama-u.ac.jp ============================================================================ ===== Date: Thu, 3 Oct 96 13:00:31 +0100 From: Konrad Hinsen To: ep7@dent.okayama-u.ac.jp Subject: Re: CHELP and CHELPG > However, the CHELP disagrees with this prediction. > That results for CHELP is old. Thus, I will recalculate the charge > on the H' with GAUSSIAN 94. Also, I will use CHELPG. You should try an SVD-based method such as CHELP-SVD. Older methods use a numerically unstable algorithm for finding the charges, and therefore might produce random numbers in difficult cases. There is a paper on this in J. Comp. Chem. 3 (1996). -- ------------------------------------------------------------------------------- Konrad Hinsen | E-Mail: hinsen@ibs.ibs.fr Laboratoire de Dynamique Moleculaire | Tel.: +33-76.88.99.28 Institut de Biologie Structurale | Fax: +33-76.88.54.94 41, av. des Martyrs | Deutsch/Esperanto/English/ 38027 Grenoble Cedex 1, France | Nederlands/Francais ------------------------------------------------------------------------------- Date: Mon, 30 Sep 1996 09:28:57 -0400 From: dew01@xray5.chem.louisville.edu (Donald E. Williams) Apparently-To: ep7@dent.okayama-u.ac.jp New Software Available Molecular interactions occur during host-substrate docking, cluster, and crystal formation - whenever molecules associate with one another. The energy and geometry of molecular association is determined by the force field. As a component of the force field, an accurate set of net atomic charges is needed. Reliable net atomic charges can be found by fitting the molecular electric potential with program PDM96. The program provides a choice of geodesic, Connolly, cubic, or user specified grid points for the electric potential. In addition to net atomic charges, program PDM96 also allows any combination of atomic dipoles/quadrupoles, bond dipoles, as well as the addition of lone pair electron sites if required. A particularly useful feature of the program is the transparent way in which fixed charges and charge dependency conditions are specified. By specififying appropriate charge dependencies (e.g., equal charges or equal sums of charges), chemical intuition can assist to produce charges which are transferable between related types of molecules. A complete error treatment with standard deviations and correlations is made. Program PDM96, Potential Derived Multipoles The following is a brief description of this program. Molecules interact with each other via their electric potential. PDM96 finds optimized net atomic charges and other site multipole representations of the molecular electric potential based on a variety of models. The program is easy to use, flexible and powerful. Results are obtained in a single iteration and a complete error treatment is made which includes estimated standard deviation and correlation of variables. The program is written in Fortran 77 and runs on any computer with F77 capability. Program PDM96 has a unique combination of features: o excellent agreement with quantum mechanical multipole moments o general sites, e.g. united atoms, not necessarily at atomic locations o each site may have any combination of monopole, dipole, or quadrupole o bond dipole model is supported o restricted (along the bond direction) bond dipole model is supported o provision for site dipole vectors in sp2 or sp3 directions o selected fixed atomic charges o selected groups of atoms with fixed charge o atomic charge equalities or symmetry relations o rotational invariance of site charges o provision for optional foreshortening of X-H bonds o comparison with Mulliken charges and Mulliken electric potential o direct input from Gaussian-92 or G-94 output file o generalized input from other quantum mechanics programs o geodesic, Connolly, and cubic grids for MEP are available o provision for custom generation of MEP grid points o error analysis with standard deviations and correlations o on-line program manual o comprehensive examples are provided A review of potential-derived charges may be found in Reviews of Computational Chemistry, Vol. II, pp. 219-271 (1991). For further information contact Dr. Donald E. Williams, Department of Chemistry, University of Louisville, Louisville, Kentucky 40292, USA. Tel:(502)852-5975 Fax:(502)852-8149 E-mail:dew01@xray5.chem.louisville.edu ----------------------------------------------------------------------- Ordering information Program package consisting of manuals, Fortran-77 source files, and demonstration example files....................................$2,000 Special discount price is available for academic use only... ........$395 Normal shipment is via ftp; please provide an account protected with a temporary password to receive the program. Inquire about shipment via other media. Make check payable to the University of Louisville and mail to: Dr. Donald E. Williams Department of Chemistry University of Louisville Louisville, KY 40292 ------------------------------------------------------------------------- I thank you for your valuable opinion. I used PDM88 (QCPE 568). The PDM88 cannot reproduce the charge on H' of H'COO-(H2O)n (n=0,1,2,3,4,.5.6) as follows: When n becomes larger, Natural population analysis predict the charge on the H' becomes more positive (J.Phys.Chem., 97, 3157(1993)). I think the prediction of Natural population analysis is correct for the following reasons: I showed the cause for the elongation of the H-C bond in HCOO- in the gas phase. This elongation results from the contribution of H-...CO2 as resonance structure to the HCOO-. When more water molecules attach HCOO-, more minus charge withdraws from HCOO- to water molecules. Consequently, it is predicted that the resonance structure, H-...CO2, contributes less to HCOO- with n increment. Due to less contribution of the resonance structure H-...CO2 with each n increment, the charge on the H' becomes more positive with n increment. However, the PDM88 disagrees with this prediction. Why PDM88 cannot reproduce the previous prediction. ============================================================================ ===== Masao Masamura Okayama University Dental School Department of Preventive Dentistry Fax: 81-86-225-3724 e-mail: ep7@dent.okayama-u.ac.jp ============================================================================ ===== From: "Donald E. Williams" Date: Thu, 3 Oct 1996 10:43:33 -0400 X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: ep7@dent.okayama-u.ac.jp Subject: Re: PDM88 Content-Type: text/plain; charset=us-ascii On Oct 3, 9:58am, ep7@dent.okayama-u.ac.jp (正村) wrote: > Subject: PDM88 > I thank you for your valuable opinion. > > I used PDM88 (QCPE 568). The PDM88 cannot reproduce the charge on > H' of H'COO-(H2O)n (n=0,1,2,3,4,.5.6) as follows: When n becomes larger, > Natural population analysis predict the charge on the H' becomes more > positive (J.Phys.Chem., 97, 3157(1993)). I think the prediction of Natural > population analysis is correct for the following reasons: I showed the > cause for the elongation of the H-C bond in HCOO- in the gas phase. This > elongation results from the contribution of H-...CO2 as resonance structure > to the HCOO-. When more water molecules attach HCOO-, more minus charge > withdraws from HCOO- to water molecules. Consequently, it is predicted that > the resonance structure, H-...CO2, contributes less to HCOO- with n > increment. Due to less contribution of the resonance structure H-...CO2 > with each n increment, the charge on the H' becomes more positive with n > increment. > However, the PDM88 disagrees with this prediction. > > Why PDM88 cannot reproduce the previous prediction. > ============================================================================ > ===== > Masao Masamura > Okayama University Dental School > Department of Preventive Dentistry > Fax: 81-86-225-3724 > e-mail: ep7@dent.okayama-u.ac.jp > ============================================================================ > ===== >-- End of excerpt from ep7@dent.okayama-u.ac.jp (正村) Dear Dr. Masamura: The fundamental question is the accuracy of the molecular electric potential calculated by your quantum mechanics program. Program pdm88 (or the improved recent release, pdm96) assumes that the MEP is accurate. In order to get an accurate MEP, you may have to use a large basis set, perhaps even including correlation. We have found, in general, that population analysis charges fit the MEP very poorly. Usually one desires the best representation of the MEP. However, there could be reasons why one would want to use non-optimal charges. Perhaps one wants to compare charges in a series of molecule with some chemical model-maybe just intuition. Or perhaps PD charges become unreasonably large or small. Pdm96 allows for input based on chemical models. For instance, fixed charges can be assigned to one or more atoms. Of course, this will degrade the fit to the MEP, but perhaps not by much. One can require charges to be equal, or for the sum of a group of charges to be zero. These "chemical" models are OK provided they do not degrade the fit to the MEP too much. Additional information about pdm96 is appended. -Donald Williams New Software Available Molecular interactions occur during host-substrate docking, cluster, and crystal formation - whenever molecules associate with one another. The energy and geometry of molecular association is determined by the force field. As a component of the force field, an accurate set of net atomic charges is needed. Reliable net atomic charges can be found by fitting the molecular electric potential with program PDM96. The program provides a choice of geodesic, Connolly, cubic, or user specified grid points for the electric potential. In addition to net atomic charges, program PDM96 also allows any combination of atomic dipoles/quadrupoles, bond dipoles, as well as the addition of lone pair electron sites if required. A particularly useful feature of the program is the transparent way in which fixed charges and charge dependency conditions are specified. By specififying appropriate charge dependencies (e.g., equal charges or equal sums of charges), chemical intuition can assist to produce charges which are transferable between related types of molecules. A complete error treatment with standard deviations and correlations is made. Program PDM96, Potential Derived Multipoles The following is a brief description of this program. Molecules interact with each other via their electric potential. PDM96 finds optimized net atomic charges and other site multipole representations of the molecular electric potential based on a variety of models. The program is easy to use, flexible and powerful. Results are obtained in a single iteration and a complete error treatment is made which includes estimated standard deviation and correlation of variables. The program is written in Fortran 77 and runs on any computer with F77 capability. Program PDM96 has a unique combination of features: o excellent agreement with quantum mechanical multipole moments o general sites, e.g. united atoms, not necessarily at atomic locations o each site may have any combination of monopole, dipole, or quadrupole o bond dipole model is supported o restricted (along the bond direction) bond dipole model is supported o provision for site dipole vectors in sp2 or sp3 directions o selected fixed atomic charges o selected groups of atoms with fixed charge o atomic charge equalities or symmetry relations o rotational invariance of site charges o provision for optional foreshortening of X-H bonds o comparison with Mulliken charges and Mulliken electric potential o direct input from Gaussian-92 or G-94 output file o generalized input from other quantum mechanics programs o geodesic, Connolly, and cubic grids for MEP are available o provision for custom generation of MEP grid points o error analysis with standard deviations and correlations o on-line program manual o comprehensive examples are provided A review of potential-derived charges may be found in Reviews of Computational Chemistry, Vol. II, pp. 219-271 (1991). For further information contact Dr. Donald E. Williams, Department of Chemistry, University of Louisville, Louisville, Kentucky 40292, USA. Tel:(502)852-5975 Fax:(502)852-8149 E-mail:dew01@xray5.chem.louisville.edu ----------------------------------------------------------------------- Ordering information Program package consisting of manuals, Fortran-77 source files, and demonstration example files....................................$2,000 Special discount price is available for academic use only... ........$395 Normal shipment is via ftp; please provide an account protected with a temporary password to receive the program. Inquire about shipment via other media. Make check payable to the University of Louisville and mail to: Dr. Donald E. Williams Department of Chemistry University of Louisville Louisville, KY 40292 ------------------------------------------------------------------------- -- Dr. Donald E. Williams email:dew01@xray5.chem.louisville.edu Department of Chemistry University of Louisville phone:502-852-5975 Louisville, KY 40292 fax: 502-852-8149 Date: Wed, 16 Oct 1996 10:17:56 -0500 From: Boyd Subject: charges and potentials To: "Masamura, Masao" X-Mailer: Mail*Link SMTP-MS 3.0.2 Content-Transfer-Encoding: 7BIT Dear Dr. Masamura, About 2 weeks ago you posted a question on the CCL asking about electrostatic potentials. Did you get all the information you needed? You can find additional information in 2 chapters of Volume 2 of "Reviews in Computational Chemistry". Donald E. Williams explains Net Atomic Charge and Multipole Models for the Ab Initio Molecular Electric Potential, and Peter Politzer and Jane S. Murray explain Molecular Electrostatic Potentials and Chemical Reactivity. In Volume 5 (1994), Steven M. Bachrach explains Population Analysis and Electron Densities from Quantum Mechanics. I hope you find these chapters helpful. Sincerely, Donald B. Boyd, Ph.D. Research Professor of Chemistry Editor, REVIEWS IN COMPUTATIONAL CHEMISTRY Department of Chemistry Indiana University-Purdue University at Indianapolis 402 North Blackford Street Indianapolis, Indiana 46202-3274, U.S.A. Telephone 317-274-6891 Facsimile 317-274-4701 Internet boyd@chem.iupui.edu REVIEWS IN COMPUTATIONAL CHEMISTRY Home Page on the World Wide Web URL http://chem.iupui.edu/~boyd/rcc.html From NMILLS@Trinity.Edu Fri Oct 25 12:32:00 1996 Received: from tucc7.tucc.trinity.edu for NMILLS@Trinity.Edu by www.ccl.net (8.8.0/950822.1) id MAA02462; Fri, 25 Oct 1996 12:26:12 -0400 (EDT) Received: from tucc2.tucc.trinity.edu by tucc7.tucc.trinity.edu with ESMTP (1.38.110.45/16.2) id AA098830666; Fri, 25 Oct 1996 11:24:26 -0500 Received: from TUCC2/MAILQUEUE by tucc2.tucc.trinity.edu (Mercury 1.20); 25 Oct 96 11:21:25 -0600 Received: from MAILQUEUE by TUCC2 (Mercury 1.20); 25 Oct 96 11:21:19 -0600 To: chemistry@www.ccl.net From: "Nancy Mills" Organization: Trinity University Date: Fri, 25 Oct 1996 11:27:56 -0500 (CDT) Subject: ZINDO/S parameters for Cl Reply-To: nmills@Trinity.Edu Priority: normal X-Mailer: Pegasus Mail/Mac v2.0.5 Message-Id: <39AFC051093@tucc2.tucc.trinity.edu> Does anyone have a set of parameters for Cl for ZINDO/S? Is there a compilation of parameters for ZINDO/S that exists? If there is interest, I will summarize to the net. Thanks. Nancy Nancy Mills Department of Chemistry Trinity University San Antonio, TX 78212-7200 210-736-7317 phone 210-736-7569 fax nmills@trinity.edu From dsmith@CTCnet.Net Fri Oct 25 13:01:59 1996 Received: from home.CTCnet.Net for dsmith@CTCnet.Net by www.ccl.net (8.8.0/950822.1) id MAA02396; Fri, 25 Oct 1996 12:00:06 -0400 (EDT) Received: from fock by home.CTCnet.Net (SMI-8.6/SMI-SVR4) id LAA13571; Fri, 25 Oct 1996 11:58:18 -0400 Date: Fri, 25 Oct 1996 11:58:18 -0400 Message-Id: <199610251558.LAA13571@home.CTCnet.Net> X-Sender: dsmith@pop.dasgroup.com X-Mailer: Windows Eudora Version 1.4.4 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: CHEMISTRY@www.ccl.net From: dsmith@CTCnet.Net (Douglas A. Smith Ph.D.) Subject: cash equivalent of supercomputer time? For anyone who owns or buys time on or operates supercomputers: What is considered the cash value of a service unit (SU) or resource unit (RU) on a Cray C90? What is considered the cash value of a service unit (SU) or resource unit (RU) on an SGI Power Challenge, say of 64 or 96 nodes or more? I will summarize. Doug -- Dr. Douglas A. Smith, President and CEO | voice: (814) 262-9091 The DASGroup, Inc. | fax: (814) 262-9337 P.O. Box 5428 | email: dsmith@dasgroup.com Johnstown, PA 15904-5428 | Contract R&D specialists in computational chemistry, process modeling, synthesis and design of novel compounds for chemistry, materials science, and biotechnology. From dsmith@CTCnet.Net Fri Oct 25 13:09:50 1996 Received: from home.CTCnet.Net for dsmith@CTCnet.Net by www.ccl.net (8.8.0/950822.1) id MAA02430; Fri, 25 Oct 1996 12:18:11 -0400 (EDT) Received: from fock by home.CTCnet.Net (SMI-8.6/SMI-SVR4) id MAA13653; Fri, 25 Oct 1996 12:16:29 -0400 Date: Fri, 25 Oct 1996 12:16:29 -0400 Message-Id: <199610251616.MAA13653@home.CTCnet.Net> X-Sender: dsmith@pop.dasgroup.com X-Mailer: Windows Eudora Version 1.4.4 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: CHEMISTRY@www.ccl.net From: dsmith@CTCnet.Net (Douglas A. Smith Ph.D.) Subject: summary of AI and NN groups On 17 October, I asked: "I am interested in finding out what academic research groups are involved in developing and using artificial intelligence, neural networks, genetic algorithms, etc. for QSAR and QSPR, for molecular modeling and property prediction, etc. Please respond directly to me; I will summarize if there is sufficient interest.I am interested in finding out what academic research groups are involved in developing and using artificial intelligence, neural networks, genetic algorithms, etc.for QSAR and QSPR, for molecular modeling and property prediction, etc. Please respond directly to me; I will summarize if there is sufficient interest." The edited summary of responses follows. Several responses either from or describing industrial/commercial work are also included. Thanks to all who answered. Doug ============================================================================ =========== The two QSAR/QSPR groups that I know of are A Hopfinger at the University of Illinois at Chicago & P. Jurs at Penn State. Both of these groups have done research in the GA and NN areas. They each have their own little niches though. If you would like some more info. on these groups, feel free to contact me -- I am quite familiar with the work from both groups. Ray Crawford Research Associate Procter & Gamble ============================================================================ =========== We are as you probably know. See our papers at 1st electronic conference of Molecular Graphics society, http://bellatrix.pcl.ox.ac.uk/mgms/ I'll be happy to provide additional information. Alexander Tropsha, Ph.D. Assistant Professor, Director the Laboratory for Molecular Modeling School of Pharmacy University of North Carolina ============================================================================ ============ I know that the department of analytical chemistry of my university applies some research in AI,NN and Gentetic algorithm, for e.g. property prediction. Their WWW-address is : http://www-cac.sci.kun.nl/cac/index-nl.html Unfortunately their home page is in dutch, however the rest is in english. So I should say look at the links research, projects, LAC verwijzingen and recente publicaties. Niels J. van der Laag Dept. of Physical Chemistry University of Nijmegen ============================================================================ =========== maybe you are interested in this URL: http://www.tu-bs.de/institute/org-chem/herges/akherges.html J.Grunenberg, Organische Chemie, Hagenring 30, 38106 Braunschweig de ============================================================================ =========== I have done a fair bit with neural nets in an academic sense (chance effects, overfitting, overtraining, pruning methods) and also apply them (when appropriate) to QSAR/QSPR problems. I can let you have references to papers and book chapters if you are interested. D.J. Livingstone ChemQuest ============================================================================ =========== We are doing nn analysis of clinical databases for database mining, disease stratification and disease process modeling using petri nets and stochastic activity nets, hypothesis evaluation against large databases, oodb development for gene discovery against clinical data, etc. Michael N. Liebman, Ph. D. Director, Bioinformatics and Genomics Vysis, Inc. ============================================================================ =========== Prof. J. Gasteiger's group at Computer-Chemistry-Center, Erlangen, Germany is one of such a research group. Please check http://www.ccc.uni-erlangen.de/ Lingran Chen, Ph.D. Chemistry Department Brandeis University ============================================================================ =========== I do a fair bit of this work for drug and crop protection agents design at CSIRO. I'm also collaborating with Prof. Frank Burden in the chemistry department, Monash University. We use NN and GAs for a number of problems in drug design, environmental and forensic chemistry. Vladimir Brusic at the Walter and Eliza Hall Institute for Medical research in Melbourne is also using NN to carry our QSAR in peptide ligands. Des Maddalena in the pharmacolgy department, University of Sydney is using NN in QSAR and clinical medicine. The groups overseas I'm involved with are run by Gerry Maggiora at Upjohn, Yoshi Takahashi at the Toyohashi University of Technology and Tariq Andrea at Du Pont's Stine Haskell Labs. Dr. David A. Winkler Principal Research Scientist CSIRO Division of Chemicals and Polymers h ttp://www.csiro.au ============================================================================ =========== Send a message to Daniel Vercauteren in Namur, Belgium: Daniel.Vercauteren@fundp.ac.be Guy Baudoux University of Namur, Belgium ============================================================================ =========== We use genetic algorithms quite a lot in our group here at Sheffield University, for a number of problems related to molecular similarity, protein structure prediction, diversity analysis, etc. A list of the projects is on our web site (www.shef.ac.uk/uni/academic/I-M/is/research/cisrg/chem.html). I use a GA for aligning electrostatic, hydrophobic and shape fields around molecules, both for QSAR and so that similarity can be calculated between them, for database similarity searching. Dr. David Wild, Department of Information Studies, University of Sheffield, Sheffield S10 2TN, U.K. ============================================================================ =========== I am working in Sheffield University, in collaboration with Glaxo Wellcome and the CCDC, on developing a genetic algorithm for flexible ligand docking. The software has recently been tested on 100 protein-ligand complexes extracted from the PDB. I've also developed a related program that superimposes small flexible molecules in order to automatically elucidate pharmacophores. This program, called GASP, is now available from Tripos Inc. If you want to know more you might want to look at these web pages. http://panizzi.shef.ac.uk/gareth/ Dr Gareth Jones Dept of Information Studies and Krebs Institute for Biomolecular Research University of Sheffield, UK ============================================================================ =========== Our research group at Moscow State University is involved in developing and using neural networks for QSAR and QSPR. We have developed in-house program NASA (Neural Approach to Structure-Activity) for Windows for this purpose. Some on-line abstracts of our papers: http://org.chem.msu.su/people/baskin/articles/ref040ab.html http://org.chem.msu.su/people/baskin/articles/ref043ab.html http://org.chem.msu.su/people/baskin/articles/ref069ft.html Dr. Igor I. Baskin Dr. Vladimir A. Palyulin ============================================================================ =========== The Institute for Tumor Biology-Cancer Research and the Austrian Research Institute for Artificial Intelligence will start a project next year. It contains the development and application of Inductive Logic Programming (ILP) methods to learn SARs for noncongeneric (carcinogenicity) databases. We will use several new approaches to include detailed biological data (species, sex and organ specific effects) and to optimize the choice of chemical descriptors. Please contact me, if you need further details. Christoph Helma Inst. f. Tumorbiologie- Krebsforschung ============================================================================ =========== You might contact Bill Dunn at U of I, Chicago. He is or was collaborating with David Rogers at MSI to devise genetic partial least squares. Presumably you also know of Johnny Gasteiger's work, extensive, in this area. Yvonne Martin ============================================================================ =========== -- Dr. Douglas A. Smith, President and CEO | voice: (814) 262-9091 The DASGroup, Inc. | fax: (814) 262-9337 P.O. Box 5428 | email: dsmith@dasgroup.com Johnstown, PA 15904-5428 | Contract R&D specialists in computational chemistry, process modeling, synthesis and design of novel compounds for chemistry, materials science, and biotechnology. From OSCAR.GONZALEZ@quimica.ulpgc.es Fri Oct 25 13:32:03 1996 Received: from procion.ulpgc.es for OSCAR.GONZALEZ@quimica.ulpgc.es by www.ccl.net (8.8.0/950822.1) id NAA02771; Fri, 25 Oct 1996 13:13:30 -0400 (EDT) Received: by procion.ulpgc.es; id AA29015; Fri, 25 Oct 1996 17:13:18 GMT Received: from ciemar4.ulpgc.es by fobos.ulpgc.es (5.65/Ultrix4.2-C) id AA25996; Fri, 25 Oct 1996 18:14:57 +0100 Received: from CIEMAR4/SpoolDir by ciemar4.ulpgc.es (Mercury 1.21); 25 Oct 96 18:13:34 GMT Received: from SpoolDir by CIEMAR4 (Mercury 1.21); 25 Oct 96 18:13:03 GMT From: "OSCAR M. GONZALEZ DIAZ" Organization: Edificio de Ciencias Basicas To: CHEMISTRY@www.ccl.net Date: Fri, 25 Oct 1996 18:12:56 GMT Subject: CaCO3 NUCLEATION, GROWING... Priority: urgent X-Mailer: Pegasus Mail for Windows (v2.10) Message-Id: <1C2AD64AAC@ciemar4.ulpgc.es> Dear netters, We are working on the nucleation and crystal growing of CaCO3. We are trying to relate induction times with thermodynamic magnitudes using Statistical Mechanics (by example..). We are looking for any references, books or helps... Many Thanks in advance Oscar Gonzalez Saludos desde Canarias, naturaleza calida... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ OSCAR GONZALEZ DIAZ DEPARTAMENTO DE QUIMICA UNIVERSIDAD DE LAS PALMAS DE GRAN CANARIA CAMPUS DE TAFIRA 35017 LAS PALMAS DE GRAN CANARIA SPAIN TLF. (34)-28-454447 FAX. (34)-28-452922 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From peng@risc.nyu.edu Fri Oct 25 14:32:06 1996 Received: from risc.nyu.edu for peng@risc.nyu.edu by www.ccl.net (8.8.0/950822.1) id NAA02918; Fri, 25 Oct 1996 13:45:08 -0400 (EDT) Received: by risc.nyu.edu (AIX 3.2/UCB 5.64/4.03) id AA31179; Fri, 25 Oct 1996 13:42:59 -0400 Date: Fri, 25 Oct 1996 13:42:58 -0400 (EDT) From: Tong Peng To: chemistry@www.ccl.net Subject: Data of amino acids? In-Reply-To: <960304125307.647@cliff.ml.wpafb.af.mil.0> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, everybody, I'd like to know where can I find data of amino acids, such their geometry, bond length and angle of peptide bonds etc. And is there an online resource of such kind of informations? Much thanks. Tong Peng peng@risc.nyu.edu From jtgolab@amoco.com Fri Oct 25 17:32:03 1996 Received: from interlock.amoco.com for jtgolab@amoco.com by www.ccl.net (8.8.0/950822.1) id RAA03804; Fri, 25 Oct 1996 17:07:50 -0400 (EDT) Received: by interlock.amoco.com id AA07848 (InterLock SMTP Gateway 3.0 for chemistry@www.ccl.net); Fri, 25 Oct 1996 16:07:49 -0500 Message-Id: <199610252107.AA07848@interlock.amoco.com> Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-3); Fri, 25 Oct 1996 16:07:49 -0500 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-2); Fri, 25 Oct 1996 16:07:49 -0500 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-1); Fri, 25 Oct 1996 16:07:49 -0500 From: jtgolab@amoco.com (Joe Golab) Date: Fri, 25 Oct 1996 16:12:12 -0500 X-Mailer: Z-Mail (3.2.2 10apr95 MediaMail) To: chemistry@www.ccl.net Subject: Fudge Factors and Computational Shortcuts Cc: jtgolab@amoco.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii We all know about the "famous" Hartree-Fock vibrational frequency fudge factor of 0.89 for calculated values. This came about from a lot of work (and observation) on simple and not so simple molecules. See Hehre, Radom, Schleyer, and Pople, "Ab Initio Molecular Orbital Theory", for example. I am wondering if there are similar "famous" fudge factors, specifically for electronic energies. For example, how much of the correlation energy does an MP2 single point calculation (at the converged HF stationary point) obtain versus the fully optimized MP2 final energy? 50%? 60%? I'm sure that for some systems there is no simple correlation since the HF description could be very poor (like O2). But does any literature exist on this subject? Word of mouth? Anecdotal? On a similar subject, is there literature (etc) on how good (if at all) hessians used in optimizations, e.g. updated, are for the calculation of the optimized structure's vibrational frequencies? Both of these questions are directed at a search for fast and quasi-reliable ESTIMATES of what a full calculation would predict. For example, at convergence, can the updated hessian resident in memory or on disk at that time be used to predict the vibrational frequencies of the stationary point (for thermochemical data) immediately? Obviously, the hessian would need to be calculated at the stationary point to get "real" answers but in the meantime, could the updated hessian's spectrum be used (for ESTIMATES) while that (longer than a diagonalization) calculation progresses? Comments will be summarized and posted in due time. -- :Joe jtgolab@amoco.com (630) 961-7878 +----------------------------------------------------------------------------+ | Every great scientific truth goes through three stages. First, people say | | it conflicts with the Bible. Next they say it had been discovered before. | | Lastly, they say they always believed it. | | - Louis Agassiz | +----------------------------------------------------------------------------+ From jsb2@camsoft.com Fri Oct 25 20:32:04 1996 Received: from camsoft.com for jsb2@camsoft.com by www.ccl.net (8.8.0/950822.1) id TAA04226; Fri, 25 Oct 1996 19:59:06 -0400 (EDT) From: Date: Fri, 25 Oct 96 19:59:00 EDT Received: by camsoft.com (4.1/3.1.090690-CambridgeSoft Corp.) id AA02998; Fri, 25 Oct 96 19:59:00 EDT Message-Id: <9610252359.AA02998@camsoft.com> To: chemistry@www.ccl.net, peng@risc.nyu.edu Subject: Re: CCL:Data of amino acids? >I'd like to know where can I find data of amino acids, such their >geometry, bond length and angle of peptide bonds etc. And is there an >online resource of such kind of informations? There's LOTS of this kind of information online. So much, in fact, you get your choice of the truth... :-) Check out http://chemfinder.camsoft.com for a cross-reference to many sites with information on all chemicals, amino acids included. Jonathan Brecher CambridgeSoft Corporation jsb@camsoft.com From shenderm@U.Arizona.EDU Fri Oct 25 21:32:04 1996 Received: from bonaire.ccit.arizona.edu for shenderm@U.Arizona.EDU by www.ccl.net (8.8.0/950822.1) id UAA04425; Fri, 25 Oct 1996 20:45:47 -0400 (EDT) Received: from localhost (shenderm@localhost) by bonaire.ccit.arizona.edu (8.7.5/8.7.3) with SMTP id RAA43084; Fri, 25 Oct 1996 17:45:22 -0700 Date: Fri, 25 Oct 1996 17:45:22 -0700 (MST) From: Mark D Shenderovich X-Sender: shenderm@bonaire.ccit.arizona.edu To: Tong Peng cc: chemistry@www.ccl.net Subject: Re: CCL:Data of amino acids? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Fri, 25 Oct 1996, Tong Peng wrote: > Hi, everybody, > > I'd like to know where can I find data of amino acids, such their > geometry, bond length and angle of peptide bonds etc. And is there an > online resource of such kind of informations? > > Much thanks. > > Tong Peng peng@risc.nyu.edu > Dear Tong Peng, The papres that introduce new MM force fields for peptides/proteins usually give a description of amino acid geometry they use. It has been done in the most systematic way by Prof. Scheraga's group in Momany et al., J.Phys.Chem. 1975, 79, 2361. For the recent corrections see also Nemethy et al., J.Phys.Chem. 1992, 96, 6472. Hope this will help you. Mark Shenderovich. From iguana@one.net Fri Oct 25 23:32:04 1996 Received: from one.net for iguana@one.net by www.ccl.net (8.8.0/950822.1) id WAA04652; Fri, 25 Oct 1996 22:47:23 -0400 (EDT) Received: from iguana (port-15-22.access.one.net [206.112.194.212]) by one.net (8.8.2/8.7.5) with SMTP id WAA23707 for ; Fri, 25 Oct 1996 22:46:04 -0400 Date: Fri, 25 Oct 1996 22:46:04 -0400 Message-Id: <199610260246.WAA23707@one.net> X-Sender: iguana@mail.one.net X-Mailer: Windows Eudora Version 2.0.3 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: CHEMISTRY@www.ccl.net From: iguana@one.net (Ray Crawford) Subject: Redundant Z-matrices All -- Does anyone out there have a little snippet of code (or know of one that is floating around) that will convert the input (or output) deck from a major software package (such as sybyl or spartan) into a redundant z-matrix which could be made into an input deck for Gaussian 94? I would prefer to find a snippet of code that I could modify to my particular need. For that matter, does anyone know of a web page (or journal article) off the top of their heads that adaquately explains the use of redundant Z-matrices with G-94? The books we have don't give a very good explaination... Any help is appreciated, Thanks, Ray Crawford iguana@one.net http://w3.one.net/~iguana "Absence of proof is not proof of absence..." -Richard Levine The Lost World Micheal Crichton