From chemistry-request@ccl.net Thu Mar 19 04:01:41 1992 Date: Thu, 19 Mar 92 17:40:54 EST From: Dave Winkler (Dr.) Subject: Alternatives to MACCS molecular database To: chemistry@ccl.net Status: R Dear netters, Does anyone know of any molecular database systems similar to MDL's MACCS database. Our application requires the storage of molecular structures and associated data such as MW, log P, pKa etc and the efficient searching of substructures and similar structures. Because of the limitations of MACCS we store the biological screening data for the MACCS structures on an Oracle database. We are having great difficulty interfacing MACCS to Oracle and there may be another competitive product which will suit our purposes better. Although we use MACCS we feel that we should be aware of any developments in the marketplace. I'm happy to summarize the responses for the net. Thanks, Dave __________________________________________________________________________ Dr. David A. Winkler Voice: 61-3-542-2244 Principal Research Scientist Fax: 61-3-543-8160 CSIRO Division of Chemicals and Polymers Private Bag 10 Clayton, Australia. From chemistry-request@ccl.net Thu Mar 19 04:04:05 1992 Date: Thu, 19 Mar 92 09:09:55 +0100 From: bio320@cvx12.inet.dkfz-heidelberg.de (Friedrich Rippmann) Subject: strand definition in DGEOM To: chemistry@ccl.net Status: R I have a few questions regarding Jeff Blaney's DGEOM Distance Geometry program. I am trying to build a IG-like beta sandwich, and I think I know where the strands are, and I know about the 'ups' and 'downs' of the side chains. Using distance constraints for the O...N distances in the sheet and a few additional distance constraints between Calphas in the two layers of the sheet gave something not too disimilar of a protein module. However, the pointing of the side chains (in, out) was, not unexpectedly, often wrong. I tried two approaches to get them in the correct orientation. First I tried to define the PHI and PSI angles, e.g. TORSION 1 17 C 1 18 N 1 18 CA 1 18 C -120. TORSION 1 18 N 1 18 CA 1 18 C 1 19 N 135. This didn't work at all. Then I imposed the appropriate PHI and PSI angles onto the input structure and tried to fix the backbone in the strands, e.g. GROUP 7 1 70 N 1 70 CA 1 70 C GROUP 7 1 71 N 1 71 CA 1 71 C GROUP 7 1 72 N 1 72 CA 1 72 C I would have expected to find the conformation of theses fixed stretches conserved, but I was wrong. Does anybody know where I missed something (the QCPE manual is somewhat sryptic)? Is there a more extensive manual or text(book) around? Thank you for any help. - Friedrich - Dr. Friedrich Rippmann Schroederstrasse 72 6900 Heidelberg 1 01049-6221-413366 c/o E.MERCK Pha Fo Chem, Drug Design 6100 Darmstadt POBox 4119 01049-6151-726290 Fax ... 710757 Email: bio320@cvx12.inet.dkfz-heidelberg.de From chemistry-request@ccl.net Thu Mar 19 06:46:38 1992 Date: Thu, 19 Mar 92 12:52:41 EET From: Jan Lundell Subject: skew angle To: chemistry@ccl.net Status: R I am interested theoretically in systems like ClHCl, ClHCl-, etc. and now I'm puzzled over skew angles, which are referred to in every I get hold on. What is the real definition and profound meaning of skew angle? Thank you for your attention and hoping to resolve this problem... Jan Lundell Dep. of Physical Chemistry University of Helsinki Helsinki, FINLAND E-mail: lundell@convex.csc.fi -- From chemistry-request@ccl.net Thu Mar 19 08:14:10 1992 Date: Thu, 19 Mar 92 12:47:42 +0100 From: martin@biokth.sunet.SE (Martin Norin, Dept. Biochem., Royal Inst. Subject: Amines-ff-parameters. To: "chemistry@ccl.net"@kth.sunet.SE Status: R Dear Netters, I have used CONCORD to generate starting structures of organic molecules for ligand-protein modelling. I'm puzzled by the amines generated by the program. An amine attached to a aliphatic chain CH3CH(NH2)CH3 is assigned to be pyramidal (sp3). However if it is attached to a aromatic ring (C6H5NH2) it is assigned to be planar (sp2). I know that the amine groups in room temperature easily may invert like an umbrella in a storm. Also that the pi-orbitals of an aromatic ring would make it more easy for the amine to become more planar. When assigning molecular mechanics force field parameters for the amine group I'm not sure how to proceed. In different biopolymer force field (AMBER, GROMOS, CHARMM) "aromatic" amines seems to be planar. I do not know the background or the history of this assignements. I suspect that an amine invovled in h-bonds may also become more planar, but this is only from my intuition and I have not done any theroretical calculations on that. My questions to you know are: 1/ What is the best assignement of a "aromatic" amine in : a) involved in a h-bond or in water solvent. b) in vacuoo or in a aprotic solvent. 2) Are there any references on this topic ) Greetings /martin ------------------------------------------------------------------ Martin Norin e-mail: martin@physchem.kth.se Dept. Biochem. KTH Stockholm Sweden. ------------------------------------------------------------------- From chiremv!andromeda!jeffb@uunet.UU.NET Thu Mar 19 12:05:10 1992 Date: Thu, 19 Mar 92 08:49:15 -0800 From: chiremv!andromeda!jeffb@uunet.UU.NET (Jeff Blaney) To: uunet!ccl.net!chemistry-request@uunet.UU.NET Subject: Re: Alternatives to MACCS molecular database Status: R You should look at Daylight Chemical Information Systems software. There are two companion programs called THOR and MERLIN which I think will handle your needs nicely without requiring you to interface to a RDB. You can reach them at: Daylight Chemical Information Systems 18500 Von Karman Ave #450 Irvine, CA 92715 (714)-476-0451 (714)-476-0654 FAX They should be able to send you demo software with a demo database containing about 5,000 structures plus the datatypes you mentioned above (and many more) for you to try out. Daylight's older software runs on VAX/VMS systems through Tektronix graphics emulation on a variety of terminals and terminal emulators. It supports both substructure and similarity searching, plus an elegant database design that allows you to store MW, log P, pKa, biological activity, or anything else you can think of - and search it. Their new software currently runs only on UNIX, with Sun and SGI systems the currently supported platforms, using a client-server model with X-windows. I've run both systems (VAX/VMS since about 1985 and the current UNIX version, which is brand new), in two very different research and corporate environments (DuPont and Chiron) and I'm convinced that Daylight's software is the best solution for my groups research needs. I've also used MDL's software quite extensively in the past; I've only seen demos of their new ISIS product - I haven't actually used it. Jeff Blaney Chiron From chemistry-request@ccl.net Thu Mar 19 12:12:00 1992 Date: Thu, 19 Mar 92 08:31:58 -0500 From: dixons%phvax.dnet@smithkline.COM Subject: RE: strand definition in DGEOM To: "chemistry@ccl.net"%INET.dnet@smithkline.com Status: R >Dr. Friedrich Rippmann: >I have a few questions regarding Jeff Blaney's DGEOM Distance Geometry program. >I am trying to build a IG-like beta sandwich, and I think I know where >the strands are, and I know about the 'ups' and 'downs' >of the side chains. Jeff reads this group and may have some further thoughts, but since I am a few time zones ahead of him, let me try to help. I don't really have enough information from your message to give you a good diagnosis of what is going wrong, but let me give you a few thoughts with regard to finding out why DGEOM doesn't do what you want at times. First of all, when you say DGEOM didn't work, do you mean the structures are not what you expected, or that you got structures that didn't obey your constraints. I suspect the later, in which case you need to look at the diagnostic messages that come out for each structure. In particular, check the report on the largest constraint violations and look at the structures that come out to see if you can see a reason for the violations. Also, check from structure to structure to see if the same distance constraints are always giving problems. Unfortunately, the error refinement step in DG tends to spread the errors out over the whole structure so it is not always easy to connect problems to specific constraints, but it is often possible to see a pattern with multiple structures. Remember that the GROUP option works by setting the upper and lower bounds of the atoms you specifiy to be equal to the corresponding distances in the input structure. In large structures, there will be many distances and so a few constraints in a rigid group don't add all that much to the overall error function. That is, it is entirely possible for the rigid group constraints to be violated since they don't count for any more than any other constraints. The same is true for TORSION constraints, since these are just translated to the corresponding distances. Also, remember that torsions and distances are not linearly related so, for some angles, fairly small distance violations result in fairly large changes in torsion angle. The default convergence criteria are set more for small molecules and are rather tight, so it is sometimes the case that, for large structures, you need to loosen the convergence criteria or use the KEEP option. Otherwise, too many structures get rejected. The is especially true when you are first working on a given problem, since it allows you to see more structures and get a better sense of what your constraints are really doing. Finally, remember that just because your distance constraints are consistent with triangle inequality checks (and hence do not generate error messages) does not mean that they are consistent with a three dimensional structure. This can mean that you have subtle problems in your constraints that can be fairly hard to track down. In general, you can see a pattern in the errors or try to find the problem by leaving out a few constraints at a time. All of these are rather general comments, since I really can't be specific about your particular problem without more information. However, they may be of general use to other DGEOM users (even if you have already thought of them all) so I will post this. >Does anybody know where I missed something (the QCPE manual is >somewhat sryptic)? >Is there a more extensive manual or text(book) around? There is really no manual for DGEOM other than the one that comes with the QCPE program. As far as textbooks, the book by Crippen and Havel (Distance Geometry and Molecular Conformation) will tell you as much or more than you ever want to know about the theory behind the distance geometry method. Good luck, Scott Dixon (dixons@smithkline.com) From chemistry-request@ccl.net Thu Mar 19 13:32:28 1992 Date: Thu, 19 Mar 1992 11:35 EST From: "DOUGLAS A. SMITH" Subject: Re: Amines-ff-parameters. To: martin@biokth.sunet.SE, chemistry@ccl.net Status: R In regards to the recent post on amine geometries: >I have used CONCORD to generate starting structures of organic molecules >for ligand-protein modelling. I'm puzzled by the amines generated by the >program. An amine attached to a aliphatic chain CH3CH(NH2)CH3 is >assigned to be pyramidal (sp3). However if it is attached to a aromatic ring >(C6H5NH2) it is assigned to be planar (sp2). > >When assigning molecular mechanics force field parameters for the >amine group I'm not sure how to proceed. In different biopolymer >force field (AMBER, GROMOS, CHARMM) "aromatic" amines seems to be planar. >I do not know the background or the history of this assignements. I >suspect that an amine invovled in h-bonds may also become more planar, >but this is only from my intuition and I have not done any theroretical >calculations on that. Lots of theoretical work has been done on aromatic amines. In fact, I have a paper in press in J. Comp. Chem. which describes some semiempirical modeling of 15 aromatic amines. We did this work because MM2(85) in MacroModel told us the amine connecting benzene and acridine rings in the topoisomerase II inhibitor m-AMSA was planar, but it also told us that the amine in aniline was planar (experimentally this latter amine is pyramidalized 37.5 degrees). Our study, which will appear in issue #5, compared MNDO, AM1 and PM3 to other theoretical methods and experiment. To make a long story short, AM1 does the best job of reproducing both experiment and ab initio data. (Our paper includes a reasonably complete literature search of this field, so may be of use for that reason as well.) 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 chemistry-request@ccl.net Thu Mar 19 14:25:28 1992 Date: Thu, 19 Mar 92 19:04:22 MET From: Axel.Eble@IMBI.Uni-Freiburg.DE (Axel Eble) Subject: Program for indexing unknown powder diffraction patterns? To: chemistry@ccl.net Status: R Howdy, I'm looking for a program to index unknown crystal/powder diffraction patternsl. I've read about shelx/86 but don't know where I could get it. Furthermore, I am interested in programs that are in the public trust - not in programs that are commercially sold. Regards, Axel P.S.: If there is enough interest, I'll summarize. -- ***************************************************************** /\ xel Eble | Email: Balrog@IMBI.Uni-Freiburg.DE / \ Bachgasse 2 | Axel.Eble@IMBI.Uni-Freiburg.DE / /\ \ 7800 Frei- | Eble@RUF.Uni-Freiburg.DE / ____ \ burg | Irc: Balrog /_/ \_\ | Smail: see above; private: see left ***************************************************************** From jkl@ccl.net Thu Mar 19 15:22:57 1992 Date: Thu, 19 Mar 1992 15:22:54 -0500 From: jkl@ccl.net To: chemistry@ccl.net Subject: Re: Good Color Postscript Printers? Status: R Forwarding to the list: ---------------------------- ==Date: Thu, 19 Mar 92 14:30:35 EST ==From: todd@sg.ab.umd.edu (Todd Tenenholz) ==To: jkl@ccl.net ==Subject: Re: Good Color Postscript Printers? I find I get fantastic results from a color slide recorder. You can cut anything on the screen out, instead of relying on the software's rendering engine, convert to Mac format, and FTP to a Mac for recording on film. In the bargain, you get a slide maker for your Mac. The quality of the image is absolutely phenomenal, and you can make prints or slides with equal ease. -------------------------------------------------------------------------------- Todd Tenenholz, Coordinator | Internet: todd@sg.ab.umd.edu | Dept. Biol. Chem. MOLECULAR GRAPHICS FACILITY | Phone: (301) 328-7165 | 660 W. Redwood St. Univ. Md. Medical School | FAX : (301) 328-8408 | Baltimore, Md 21201 -------------------------------------------------------------------------------- From PEARLMAN@VAX.PHR.UTEXAS.EDU Thu Mar 19 16:11:07 1992 Date: Thu, 19 Mar 1992 15:13:22 -0600 (CST) From: PEARLMAN@VAX.PHR.UTEXAS.EDU Subject: RE: Amines-ff-parameters. To: chemistry-request@ccl.net Status: R Martin Norin posted the following CONCORD-related question: > I have used CONCORD to generate starting structures of organic molecules > for ligand-protein modelling. I'm puzzled by the amines generated by the > program. An amine attached to a aliphatic chain CH3CH(NH2)CH3 is > assigned to be pyramidal (sp3). However if it is attached to a aromatic ring > (C6H5NH2) it is assigned to be planar (sp2). > > I know that the amine groups in room temperature easily may invert like an > umbrella in a storm. Also that the pi-orbitals of an aromatic ring would > make it more easy for the amine to become more planar. ... I > suspect that an amine invovled in h-bonds may also become more planar, > but this is only from my intuition and I have not done any theroretical > calculations on that. > > What is the best assignement of a "aromatic" amine in : > > a) involved in a h-bond or in water solvent. > b) in vacuoo or in a aprotic solvent. The height-above-the-plane distance for the nitrogen in NH3 is approximately 0.33A. I have performed geometry optimizations on C6H5NH2 using both semi- empirical and ab initio MO methods and found height-above-the-plane distances for the nitrogen ranging from 0.09 to 0.18A depending on hamiltonian and basis set. I have also performed geometry optimizations on C6H5NH2--H2O complexes (although only using semi-empirical methods) and found height-above-the-plane distances ranging from 0.05 to 0.18 depending on hamiltonian and complex geometry. Clearly, neither true sp2 (planar, 120 angles) nor true sp3 (tetrahedral, 109.47 angles) is the best representation of the aniline geometry; an intermediate "hybridization" would be better. Although CONCORD is NOT a molecular mechanics program, it (like MM programs) must choose either sp2 or sp3. After considerable thought and discussion with CONCORD users, I chose sp2. The main reason for this choice was convenience. Since CONCORD- generated structures are often used in a 3D-searching or "automated-docking" context, the sp2 choice avoids some problems (or extra work) associated with the assymetry of a non-planar structure. While I was writing this message, I received and read with interest the message from Doug Smith. His forthcoming paper on the geometry of aromatic amines should be quite interesting. He also commented that "MM2(85) in MacroModel ... told us that aniline is planar." This is entirely the result of MacroModel assigning an sp2 atom-type to the nitrogen. In fairness to Lou Allinger we should note that stand-alone MM2(85) tells us that "it doesn't have the parameters needed to do aniline correctly!" :-) Regards, Bob Pearlman From chemistry-request@ccl.net Thu Mar 19 17:55:57 1992 Date: Thu, 19 Mar 92 17:02:30 PRT From: Dionisia Maria de Jesus Subject: Hi Postgraduate info To: CHEMISTRY@ccl.net Status: R I finished the chemistry course. I'm searching for a postgraduate info (University with this course, scien tific investigation and more information about this subject) My area is in physical chemistry or teoric chemistry. Thanks in advance Dionisia Maria de Jesus E-mail: Bitnet : sceioc23@ptearn Internet : sceioc23%ptearn.bitnet@cunyvm.cuny.edu Address: Dionisia Maria de Jesus Campo Grande 4 - 4 DIREITO 1700 Lisboa Portugal From chiremv!andromeda!jeffb@uunet.UU.NET Thu Mar 19 20:27:01 1992 Date: Thu, 19 Mar 92 16:43:01 -0800 From: chiremv!andromeda!jeffb@uunet.UU.NET (Jeff Blaney) To: uunet!ccl.net!chemistry-request@uunet.UU.NET Subject: Re: strand definition in DGEOM Status: R The only additional suggestions I have beyond Scott Dixon's comments are: 1) Can you include CB-CB lower/upper bounds between residues on opposing strands? These distances should be sensitive to whether the two sidechains are both up, both down, or pointing opposite directions. 2) Chiral constraints can be included on any 4 atoms; they don't all have to be attached to the same carbon. Without knowing more about what you're trying to build I can only guess, but I think you might be able to define a chiral constraint (which is actually a signed volume enclosing 4 atoms) between groups of 4 CB's at a time, since you say you know which ones should point up and down. By using the CALC option on the CHIRAL constraint you could determine the approximate chiral volume defined by any 4 CB's, and then apply a chiral constraint with a delta value (e.g. chiral +/- delta). I hope this helps. I apologize for the cryptic documentation; I'm working on a new release of DGEOM and I'll try to add more detail and hints to the manual (including Scott's comments). A few time zones later, Jeff Blaney Chiron From chemistry-request@ccl.net Thu Mar 19 20:47:15 1992 Date: Thu, 19 Mar 1992 18:30 CST From: Andy Holder Subject: "our sacred honor..." To: CHEMISTRY@ccl.net Status: R Greetings Netters. All of us of a serious computational chemistry persuasion have tried at one time or another to "sell" our utility and that of computational procedures to others (doubters, people who do actual EXPERIMENTS!!) I am particularly interested in sem- empirical methodology (what a great big surprise) and I was wondering about the success stories out there. I have found very few in the literature, but I am sure that they exist. I know that many of them are corporate proprietary and you can't reveal them. (By the way, did you know that there are a lot of corporate people that subscribe to the list that are not allowed to post by their companies? Surprised me.) So please, if there is something in the literature or something you can tell me about that fits in here please respond directly. I'll be happy to compile and repost when all added up. I think that this will be a valuable resource for all of us and I hope that you will contribute. Perhaps another kind soul would or- ganize similar efforts for ab initio and the molecular mechanics/ dynamics people. Andy Holder =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= DR. ANDREW HOLDER Assistant Professor of Computational/Organic Chemistry Department of Chemistry || BITNET Addr: AHOLDER@UMKCVAX1 University of Missouri - Kansas City || Internet Addr: aholder@vax1.umkc.edu Spencer Chemistry, Room 502 || Phone Number: (816) 235-2293 Kansas City, Missouri 64110 || FAX Number: (816) 235-1717 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From chemistry-request@ccl.net Thu Mar 19 23:07:34 1992 Date: Thu, 19 Mar 1992 21:42 EST From: "DOUGLAS A. SMITH" Subject: MacroModel Aniline structure To: chemistry@ccl.net Status: R One quick qualification regarding Bob Pearlman's comments. While I did state that MacroModel modeling was done using the MM2(85) parameters, I made a significant and glaring omission. MacroModel has and uses a special substructure parameter set for aniline (many others are available as well for other molecules). So the blame is not Lou Allinger's, nor is it Clark Still's. I don't want to assign blame at all, just note that molecular mechanics didn't get it right. Doug Smith reference for aniline substructure parameters derived by WCS from: for stretching, bending and torsions of the aromatic ring not involving the nitrogen atom, the benzenoid substructure was used - H.-D. Beckhaus, Chem. Ber., 1983, 116, 86. for torsions involving the anilino nitrogen, the aniline parameters used were derived from T. B. Grindley, A. R. Katritzky, R. D. Topsom, J. Chem. Soc. Perkin Trans. II, 1974, 289.