From shin@nc.u-tokyo.ac.jp Wed Aug 21 05:49:38 1991 To: jkl@ccl.net Subject: Re: What Crystallographic Software is available? Date: Wed, 21 Aug 91 18:49:50 +0900 From: Shin Yoshimura Status: R >> >I am interested in information concerning >> >all commercial and public domain softwares >> >available for Crystal Structure solution >> >for both small and large molecules. We are now using XTAL, which is total package for Crystal Structure analysis including interactive graphics. The graphics of XTAL is based on GKS (SunGKS, xgks, and others), phigs version is now developing, I hear. XTAL is distributed from Western Australia University, and it costs $500 for academic or $2000 for commercial. Contact: Prof. Syd Hall Shin Yoshimura The University of Tokyo From chemistry-request@ccl.net Wed Aug 21 10:05:03 1991 Date: Wed, 21 Aug 1991 09:47 EDT From: "DOUGLAS A. SMITH" Subject: AMPAC 2.1 Docs To: chemistry@ccl.net Status: R I just retrieved AMPAC 2.1 from UT-CHEM.CM.UTEXAS.EDU via anonymous ftp. There were no doc files. Does anyone know where I can get a copy of the documentation on line? I am installing the program on several UNIX machines. Doug Smith Assistant Professor of Chemistry The University of Toledo Toledo, OH 43606-3390 voice 419-537-2116 fax 419-537-4033 email fax0236@uoft02.utoledo.edu From jkl@ccl.net Wed Aug 21 12:32:56 1991 Date: Wed, 21 Aug 91 11:48:20 EDT From: jkl@ccl.net Subject: Reminder, To: chemistry@ccl.net Status: R Judging by the type of messages which appear recently on the list, it seems the reminder is due. 1. You send to CHEMISTRY@ccl.net only messages to be received by all 670+ subscribers to the list !!! 2. All other messages you send to CHEMISTRY-REQUEST@ccl.net or JKL@ccl.net or JKL@OHSTPY.BITNET Please, remember: CHEMISTRY is all 670+ people, CHEMISTRY-REQUEST is if when you request something. I urge you to read the help file for the list. You can get it by sending one line message: send help from chemistry to OSCPOST@ccl.net or OSCPOST@OHSTPY.BITNET Jan Labanowski The List Coordinator jkl@ccl.net, JKL@OHSTPY.BITNET From Leif.Laaksonen@vtt.fi Wed Aug 21 14:10:32 1991 Date: Wed, 21 Aug 91 21:11:54 GMT From: Leif.Laaksonen@vtt.fi (Leif Laaksonen) To: chemistry-request@ccl.net Status: R Can somebody give me a pointer to how one fits two molecular structures on each other. I can well understand how one calculates the translation leading to the smallest least square but how do I calculate the rotation? It would of course be just great to receive the piece of code for doing the rotation. Are there any other approaches for this problem? Thanks in advance, Leif Laaksonen **************************************************************************** Leif Laaksonen I "Learn from yesterday, Phone: 358-0-4565105 I live for today, Fax: 358-0-4552028 I look to tomorrow, I rest this afternoon." Mail: laaksone@geeni.bio.vtt.fi I -- Snoopy or I laaksonen@finfun.bitnet I --------I-------- Post address: I Street address: VTT/Biotechnical laboratory I VTT/Bio P.O. Box 202 I Tietotie 2 SF-02151 ESPOO I ESPOO FINLAND I **************************************************************************** From chemistry-request@ccl.net Wed Aug 21 14:12:01 1991 Date: Wed, 21 Aug 91 13:41:19 EDT From: R2VJV%AKRONVM@vm1.cc.UAKRON.EDU Subject: IBM-6000 ... To: chemistry@ccl.net Status: R hi folks, We are in the process of buying a workstation/IBM-6000 machine in our group for computations like Molecular Dynamics/Monte Carlo of chain molecular f luids. Our current works relate to developing theories for thermodynamics of ch ain molecular fluid mixtures and comparing them to simulation results. We are a lso developing MD routines for hydrogen bonding in fluid mixtures. We have been using our own routines on IBM-3090 and Cray Y-MP at OSC. We are interested in buying a commercial software like polygen for our computations. I am not sure whether Polygen is transportable to IBM-6000. Or have they already got them running on IBM-6000. My question to the 'list' is regarding the advant ages of buying a IBM-6000 machine compared to a workstation for the same cost. Do IBm-6000 handle sophisticated graphics like workstations. Do they have graphics software available for ball-stick modelling. I would appreciate if yo u could share with me your experiences regarding this. We want to invest right. Thanks in advance. venkat From djh Wed Aug 21 15:03:41 1991 Date: Wed, 21 Aug 91 14:54:47 -0400 From: djh@ccl.net To: chemistry@ccl.net Subject: fitting molecular structures Status: R Leif.Laaksonen@vtt.fi (Leif Laaksonen) writes: >Can somebody give me a pointer to how one fits two molecular structures >on each other. I can well understand how one calculates the translation >leading to the smallest least square but how do I calculate the rotation? >It would of course be just great to receive the piece of code for doing >the rotation. Are there any other approaches for this problem? A FORTRAN code of mine (fitest.f) for fitting rigid structures is available via anonymous ftp on www.ccl.net, in the pub/chemistry/fitest directory. It uses quaternions to effect the rotation and is not iterative (except for matrix diagonalization). Fitest should compile and load on a VAX, Sun, Iris, or CRAY. Most of the file is garbage: a main program and subroutines to generate a random reference and test molecule and fit them. The useful stuff is the routine qtrfit. It should be self-explanatory (ha!). I haven't been able to extend the method to work with a series of molecules. -- David J. Heisterberg djh@ccl.net We are NOT all The Ohio Supercomputer Center djh@ohstpy.bitnet Keynesians now. Columbus, Ohio 43212 ohstpy::djh From jkl@ccl.net Wed Aug 21 15:35:38 1991 Date: Wed, 21 Aug 91 15:14:34 EDT From: jkl@ccl.net Subject: fitest.f To: chemistry@ccl.net Status: R If you do not have ftp, and want a copy of the fitest.f file you can send the following, strange-looking message: send fitest/fitest.f from chemistry to OSCPOST@ccl.net or OSCPOST@OHSTPY.bitnet, and the file (circa 19kB) will be automatically forwarded to you. Jan jkl@ccl.net From chemistry-request@ccl.net Wed Aug 21 16:13:19 1991 Date: 21 August 1991 14:43:46 CDT From: "Bob Goldstein 6-6664" To: Subject: Status: R About fitting two molecular structures to each other: see Kabsch in Acta Cryst. A32 (1976) 922-923. Basically, substract the coordinates, atom pair by atom pair, then square and sum up to get a chi square. Then adjust the rotation and translation to minimize this chi-square. If you take as variables each element of the rotation matrix and the translation vector, subject to the constraint that the rotation matrix must be orthogonal, it turns out that the minimization is actually linear in the variables, so can be solved analytically. The derivation (with my own small twist involving SVD) is briefly: G = goodness of fit (chi square) A_i = 3 cartesian coordinates of atom i in molecule A B_i = 3 cartesian coordinates of atom i in molecule B R = 3X3 rotation matrix T = translation vector C_a = center molecule A The objective is to translate and rotate molecule B to minimize G, where G = sum_i [ A_i - C_a - R (B_i - T) ]^2 Use Lagrange multipliers to ensure that R remains orthogonal, and then minimize G with respect to each element of R and T individually. The result is T = sum_i B_i (not too surprising) and M = R S, where M = sum_i [ A_i -C_a] [B_i - T], and S = some symmetric matrix involving the Lagrange multipliers. Thus you must form M, and then decompose it into the product of an orthogonal and symmetric matrix. The resulting orthogonal matrix is the rotation matrix of interest. Kabsch solves this problem by squaring M, but I note that singular value decomposition allows M to be written M = U W V^T where W is diagonal, and U and V^T (V transpose) are orthogonal since M is square. But then M = [U V^T ][ V W V^T] and therefore U V^T is exactly the rotation matrix we desire: so do the SVD, then set the singular values all to one, and remultiply the factors to get the rotation matrix! The only caveat is if one or more singular values are zero (which would happen if the molecules happened to be planar). In this case, the resulting orthogonal matrix might actually contain a reflection in addition to a rotation. Well, enough droning on. I have some fortran to do this, if anyone wants. bobg From chemistry-request@ccl.net Wed Aug 21 17:06:27 1991 Date: Wed, 21 Aug 1991 13:55 MST From: YAMAMURA%CGF@Arizona.edu Subject: Summary of Circular Dichroism, Secondary Structure Determination Prgms To: chemistry@ccl.net Status: R Thanks to the many people who responded to my request for Circular Dichroism information, it is now possible for me to post a summary of the information received: ________________________________________________________________________ 1. From Lisa M. Balbes, Ph.D. Research Triangle Institute P.O. Box 12194 Research Triangle Park, NC 27709-2194 balbes@osiris.rti.org 919-541-6563 work 919-490-1888 home During graduate school, I spent a significant amount of time evaluating a variety of Cd algorithms. I did an extensive evaluation of PROSEC, NCOMP, and VARSELEC. If you would like hard copy (or a mac disk) with the chapters of my dissertation with my results, send me a USmail address. I found VARSELEC to be by far the best -- by Curtis Johnson. He gave us a PC version, we modified if for the VAX. If you decide to use VARSELEC I also have some instructions that I wrote to teach others. Some references: PROSEC - supplied with the AVIV spectrometer, see Anal. Biochem (1978) 91, 13-31. NCOMP - unpublished algorithm of M.S. Briggs and A. N. Stroup, University of Delaware -- contact Lila Gierasch (Texas) VARSELEC - Biochemistry (1981) 20, 1085-1094. Anal. Bichem. (1987) 67, 76-85. Ann. Rev. Biophys. Biophys. Chem (1988) 17, 145-166. (and lots more) ________________________________________________________________________________ >From Dr. Wally E. Reiher II Tripos Associates, Inc. P.O. Box 485 Pasadena, CA 91102 wer@tripos.com 213-259-8001 Article evaluating CONTIN (negatively) A. Bobba, et al., "Estimation of protein secondary structure from circular dichroism spectra: a critical examination of the CONTIN program", Protein Seq Data Anal (1990) 3: 7-10. cited references for CONTIN: S.W. Provencher, CONTIN Users Manual, EMBL Technical Report DA05, 1982. European Molecular Biology Laboratory, Heidelberg ______________________________________________________________________________ Try VARSELECT. It is written in fortran, source is included. IBM high density disks are available for negligible cost. Uses arbetrary sized basis set (twenty supplied). Single value decomposition algorithm. The limitation is that the unknown must be collected with the same point spacing as the basis set data. I have a version in which I added a spline fit to make the increment size self adapting -- but you should contact the authors of the program first. Araz Toumadje Dept. Biochemistry and Biophysics Oregon State University Weniger Hall 535 Corvallis, OR 97331 Contact me be E-mail at kxb3@po.cwru.edu Ken Butenhof, CWRU Cleveland OH -- Ken Butenhof kxb3@po.cwru.edu Summary: It appears that VARSELECT is a popular method for Circular Dichroism Studies and that some recent improvements have been made to the program. Those who are interested in VARSELECT should probably contact Dr. Curtis Johnson and Araz Toumadje at Oregon State University. I thank all of those who contributed their time and information in responding to our request. Susan Yamamura yamamura@cgf.chem.arizona.edu University of Arizona Chemistry Department From chemistry-request@ccl.net Wed Aug 21 17:51:58 1991 Date: Wed, 21 Aug 91 17:19:33 EDT From: enrico@watson.ibm.com To: chemistry@ccl.net Status: R > ............................ My question to the 'list' is regarding the advant >ages of buying a IBM-6000 machine compared to a workstation for the same cost. >Do IBM-6000 handle sophisticated graphics like workstations. Do they have > graphics software available for ball-stick modelling. ................. The IBM RISC System/6000 "is" a workstation. The models available go >from a 20MHz, 8MB to 32MB of memory and 160MB to 640MB of disk space workstation (Model 320) to a 41MHz, 64MB to 512MB of memory and 800MB to 2.5GB of disk space workstation (Model 550). There is also a special model for high speed graphics (Model 730) and a model for using as a high capacity file server (Model 950). The graphics capabilities of these machines include 8-bit and 24-bit color adapters that support 1280x1024 displays. There are also special graphics processors that can be attached to the workstations (GTO, GT4X, GT4, GT3) that provide high speed graphics capabilities similar to the Model 730. All support PHIGS and some of them also IRIS GL. There is also a special Graphics Visualization System that attaches to a workstation in three models that have from 8 processors and 256 MB of memory to 32 processors and 768 MB of memory (more memory is possible) for ultimate performance in 3D graphics. The workstations run IBM AIX Version 3 that is an IBM version of UNIX. There are IBM versions of X-Windows, PHIGS, FORTRAN, Pascal, C, etc., etc., etc. I don't know about other programs, but our group distributes the MOTECC package that contains KGNGRAF. This is a program for displaying molecules using balls and sticks, or van der Waals spheres, or Dreidig-like models. It also has a utility for building proteins. It can produce 3D displays of electronic density, etc. There are many more features. Hope this helps, Omar G. Stradella From chemistry-request@ccl.net Wed Aug 21 18:33:36 1991 Date: Wed, 21 Aug 91 17:17:26 CDT From: erickson@k.scs.uiuc.edu (Jon Erickson) To: chemistry@ccl.net Subject: Need code to compute Kinetic isotope values from Hartree-Fock frequency data. Status: R I'm interested in calculating kinetic isotope effect values >from Hartree-Fock frequency data, but I don't have the code to carry these calculations out. I have the transition states and reactant geometries and can obtain the deuterium substiuted analogs. Does anyone know of any available code to calculate the KIE (ie kH/kD) values. If the code is compatible with standard ab initio MO packages (Gaussian, CadPac, Spartan) these will be the most useful to me. Thanx in advance for any help. Jon Erickson Department of Chemistry University of Illinois at Urbana-Champaign 471 Roger Adams Laboratory (Box 73-5) 1209 West California Street Urbana, IL 61801 USA Phone: (217) 333-3509 work (217) 384-7940 home Internet, CSNET: erickson@k.scs.uiuc.edu BITNET: erickson@uiucscsk