From anthony.scott@anu.edu.au Mon Sep 26 20:00:41 1994 Received: from anu.anu.edu.au for anthony.scott@anu.edu.au by www.ccl.net (8.6.9/930601.1506) id TAA26378; Mon, 26 Sep 1994 19:11:00 -0400 Received: from cscgpo.anu.edu.au by anu.anu.edu.au (4.1/SMI-4.1) id AA16986; Tue, 27 Sep 94 09:10:30 EST Received: from huxley.anu.edu.au by cscgpo.anu.edu.au (4.1/SMI-4.1) id AA13700; Tue, 27 Sep 94 09:12:56 EST Date: Tue, 27 Sep 94 09:12:56 EST From: anthony.scott@anu.edu.au (Anthony P Scott) Message-Id: <9409262312.AA13700@cscgpo.anu.edu.au> To: chemistry@ccl.net, cmartin@rainbow.uchicago.edu Subject: Re: CCL:Questions about charges Dear Charles, You might look at a recent paper by Ken Wiberg: J. Comp. Chem. 14, 1504, (1993) in which he examines such issues. Kind regards, Anthony P. Scott Research Officer Computational Chemistry Group Research School of Chemistry Australian National University Canberra, ACT, Australia From BAELL@mel.dah.csiro.au Mon Sep 26 21:00:42 1994 Received: from dahsun.mel.dah.csiro.au for BAELL@mel.dah.csiro.au by www.ccl.net (8.6.9/930601.1506) id UAA26971; Mon, 26 Sep 1994 20:17:55 -0400 Received: from mhs.mel.dah.CSIRO.AU by dahsun.mel.dah.csiro.au with SMTP id AA22492 (5.67a/IDA-1.5 for <@dahsun.mel.dah.csiro.au:chemistry@ccl.net>); Tue, 27 Sep 1994 10:20:34 +1000 Message-Id: <199409270020.AA22492@dahsun.mel.dah.csiro.au> From: BAELL@mel.dah.csiro.au (Jonathan Baell) To: chemistry@ccl.net (chemistry) Subject: atomic charges Date: Tue, 27 Sep 94 10:15 Chuck I believe the ESP subroutine of MOPAC 6 uses the "deorthogonalization" technique of Besler, Merz and Kollman (J.Comp.Chem. 11, 431-439 (1990)). They found that it was "possible to obtain 6-31G* quality point charges by simply scaling the MNDO ESP charges", as the correlation between the two was very good. The introduction of this paper covers a lot of background/history on electrostatic calculations. A quasi-recent issue of J. Comp.-Aided Mol. Design (vol. 5 (1991)) was devoted to electrostatics, and a future issue was then envisaged, but I haven't kept up with it. I didn't see Feng Zhou's message: if you have the email address, perhaps you could forward this message? Jonathan Ball +----------------------------------------------------------------------+ Dr Jonathan Ball Senior Research Scientist CSIRO, Division of Animal Health Private Bag 1, Parkville, Victoria 3052, Australia Internet email: baell@mel.dah.csiro.au Tel: +61 3 342-9782 Fax: +61 3 347-4042 +______________________________________________________________________+ From nash@chem.wisc.edu Mon Sep 26 22:00:44 1994 Received: from fozzie.chem.wisc.edu for nash@chem.wisc.edu by www.ccl.net (8.6.9/930601.1506) id VAA28373; Mon, 26 Sep 1994 21:46:11 -0400 Received: by fozzie.chem.wisc.edu; id AA11309; 5.57/42; Mon, 26 Sep 94 18:24:56 -0500 Message-Id: <9409262324.AA11309@fozzie.chem.wisc.edu> X-Sender: nash@fozzie.chem.wisc.edu Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Mon, 26 Sep 1994 18:24:43 -0500 To: chemistry@ccl.net From: nash@chem.wisc.edu (John R. Nash) Subject: Natural Charges As has been mentioned, the Mulliken population method is not the only way to get atomic charges from an ab initio method, and is often of dubious significance. I thought I'd point out that GAUSSIAN92 easily allows the calculation of "natural" charges with the POP=NBO keyword. (See the manual for the literature refs; I don't have it handy.) Again, this is one of several methods -- be sure to understand how charges are arrived at before making interpretations of them. -===-John R. Nash-==-nash@chem.wisc.edu-==-UW-Madison Chem. Dept-===- From ross@cgl.ucsf.EDU Mon Sep 26 22:08:18 1994 Received: from socrates.ucsf.EDU for ross@cgl.ucsf.EDU by www.ccl.net (8.6.9/930601.1506) id VAA28095; Mon, 26 Sep 1994 21:30:39 -0400 Received: (from ross@localhost) by socrates.ucsf.EDU (8.6.9/GSC4.24) id SAA12233; Mon, 26 Sep 1994 18:30:29 -0700 Date: Mon, 26 Sep 1994 18:30:29 -0700 Message-Id: <199409270130.SAA12233@socrates.ucsf.EDU> From: ross@cgl.ucsf.edu (Bill Ross ) To: chemistry@ccl.net Subject: 'raw speed' Cc: ross@cgl.ucsf.EDU Richard Walsh of Minnesota Supercomputer Center, Inc. writes: I am not sure what you mean by 'raw speed' here ... could you also report C90 parallel times or those for the T3D? It would be helpful and more complete to present these comparisons. I certainly should have defined 'raw speed' - it means the best you can get out of the code we distribute. I look forward to being able to present such comparisons; we should have T3D results at least for the next release of Amber, and I'm hoping for C90 parallelization too. The parallel Charmm Web page is the best I have to offer for now along these lines. My guess is that your benchmark would run with the most 'raw speed' on a C90 16/512 with a T3D attached. I chopped a discussion of possibilities out of my post in the interest of sticking to data. In support of Richard's notion, I might add that I believe that even single-cpu Amber could be optimised further on Crays to get better vector performance and maybe beat these results (the nonbond loop only handles 2 residues at a time, so the pipeline setup cost overwhelms the vector saving on solvated systems and breaks even in vacuum). Other possibilities for the next record time include message-passing on IBM's SP2, Intel's Paragon, or on a cluster of multi-cpu TFP's. when does a workstation become a supercomputer? ... when it costs $500,000.00, $1,000,000.00. The new SGI systems with multiple CPUs and functionally configured can easily exceed these numbers. Around here anyway a workstation is something that can be purchased for about $50,000. Good point. My unconscious assumption was that the thing with a main graphics screen on someone's desk that doesn't need a machine room and site engineer(s) to service it is a workstation, and I haven't thought so much about price. The SGI TFP performance numbers are impressive, but they should be presented here with as high a regard for completeness as for enthusiasm (excepting perhaps by those selling the hardware involved :-)). Hopefully this will have allayed your worries. I try to be fair by reporting each new 'speed record' for our code as it happens. (No one complained about the earlier announcement about Fujitsu followed by the C90 one.) I agree there is some arbitrariness involved in terms of which machines get optimised when. In this case SGI worked on shared memory parallelism; Cray could have but didn't choose to, though there has been talk of them doing it for the next release. My hope is that, on the whole, these benchmarks provide a relatively unbiased means of comparing machines. By the way, I expect that the benchmarks for the next release of Amber will eliminate some redundancies and more usefully reflect the molecular mechanics regimes one might choose, at least by using different cutoff sizes. Bill Ross University of California, San Francisco From jeremy@med.su.oz.au Mon Sep 26 22:11:52 1994 Received: from blackburn.med.su.oz.au for jeremy@med.su.oz.au by www.ccl.net (8.6.9/930601.1506) id VAA27839; Mon, 26 Sep 1994 21:10:02 -0400 Received: (jeremy@localhost) by blackburn.med.su.oz.au (8.6.8.1/8.6.4) id LAA13120 for chemistry@ccl.net Tue, 27 Sep 1994 11:09:59 +1000 From: Jeremy R Greenwood Message-Id: <199409270109.LAA13120@blackburn.med.su.oz.au> Subject: Gaussian92 on 486 summary To: chemistry@ccl.net Date: Tue, 27 Sep 1994 11:09:58 +1000 (EST) X-Mailer: ELM [version 2.4 PL23] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 7119 In response to my query regarding run times and memory requirements for running Gaussian92 on a 486, I received the following replies: ---------------------------------------------------------------------------- From: ecoer@sb615.rivm.nl Subject: CCL:Gaussian '92 on a 486 -Reply Hi Jeremy, Last month I have installed G92 for Windows on my 486/66MHz with 16MB internal memory and so far I have only tried to do some normal and transition state optimization jobs on chloro-methanes/ethanes. But run-times are rather lengthy I'm afraid. For tetra-chloromethane one SCF at the HF/6-31+G(d) level took over 5 hours (about 260 gaussians). Using over 12 heavy atoms will take days I think. I have 200Mbyte of space on my hard disk but I suppose less will also do. I am very interested in the reactions you receive, because I am also still figuring out what will be possible in terms of system size. Please summarize to the net or to me Thanks in advance, Emiel Rorije / E.Rorije@rivm.nl / ecoer@rivm.nl Laboratory for Ecotoxicology National Institute of Public Health and Environmental Protection The Netherlands. ------------------------------------------------------------------------------ From: danne@rschp1.anu.edu.au (Danne R Rasmussen) Subject: Re: Gaussian '92 on a 486 Status: OR Jeremy, How is it all going? I saw your CCL posting and thought I'd drop you a little line. I think for a start that turbomol would be better for running HF and MP2 jobs than GAUSSIAN. We have the source for turbomol so you could install a public domain unix (say linux) on your 486/Pentium? and try to compile turbomol under that system (shouldn't be too hard to get up and running). As for the amount of disk turbomol again is much more efficient than GAUSSIAN at running semidirect/conventional MP2. Fulldirect MP2 uses very little disk but needs heaps of core memory (more expensive) and in general is slower (because you have to recompute the integrals all the time) except on vector machines like the vp (where the time to compute the integrals is negligible). As you know G92's semidirect MP2 is tunable as to how much disk you need for optimisations but fixed for frequencies (unless you do them numerically). I find that for C13H20 optimisations at MP2/6-31G* with G92 in high symmetry (D2) I need 2Gb of disk (in order to do the MP2 in 3 passes). This is really pushing the limits of a workstation (RS/6000 m355 with 64Mb). For the slightly smaller job C9H12 MP2/6-31G* in lower symmetry (C2) I needed only 1Gb of disk to get through the MP2 in a single pass and it ran in only 20Mb of core. Hope this helps, ______________________________________________________________________________ Danne R Rasmussen Phone: +61 6 249-3771 Research School of Chemistry +61 6 249-2018 Australian National University Canberra, ACT 0200 Fax: +61 6 249-0750 AUSTRALIA E-mail: danne@rsc.anu.edu.au ------------------------------------------------------------------------------ From: csonka@iris.inc.bme.hu (Csonka Gabor) Subject: Re: CCL:Gaussian '92 on a 486 Dear Jeremy Greenwood, my estimation for a 20 atom molecule with MP2/6-31g* level of theory is from 2 to 6 months on a 486/33 MHz. You need at lest 2 GB disk for that to run confortably. Symmetry can help a lot the 2 month is valid if you have some symmetry. I did MP2/6-31G* geometry optimization for 24 atoms (11 heavy) with c3 symmetry on SGI Iris 4000SC 100MHz. One step took about 6 days. PC is 10 to 15 times slower with g92. Maybe a 100 MHz pentium can do the job, but as far as I know g92 is not optimized for pentium, so I expect a rather moderate performance increase. The MP2 uses lot of disk I/O which are slow on PC-s. You need the fastest available disk. If you add all the cost you will easily reach the price of a unix workstation. ---------------------------------------------------------- Gabor I. Csonka Budapest University of Technology Tel/FAX: (361) 18.12.177 Inorganic Chemistry Dept. Ch. Bldg csonka@iris.inc.bme.hu H-1111, Bp. Szent Gellert ter 4 ---------------------------------------------------------- From: "Gregory L. Durst - DowElanco R&D" Subject: RE:CCL:Gaussian '92 on a 486 Jeremy, I would recommend you look at the book "Exploring Chemistry with Electronic Structure Methods: A Guide to Using Gaussian", by JB Foresman & A Frisch, Gaussian Inc., 1993. It's $35, (well...last summer it was $35) and has some timings for the test suite of molecules on a Multiflow TRACE 14/300 computer with 64 MB RAM. At the bottom of the table they say that a 50 MHz 486 should take roughly 10-15 times longer. (eg. Ethylene geom optimization, tutorial #009, used 3-21G basis set, and took 2 min 17.7 secs on the TRACE, so figure 20-30 minutes on the 486.) Hope this helps! Greg +-----------------------------------------------------------------------+ | Gregory L. Durst Computational Chemistry | | phone: 317/337-3413 DowElanco R&D | | email: gdurst@dowelanco.com 9330 Zionsville Rd. Bldg 306/D2 | | Indianapolis, IN 46268 USA | +-----------------------------------------------------------------------+ --------------------------------------------------------------------------- From: WILLSD@conrad.appstate.edu Subject: CCL:Gaussian '92 on a 486 (fwd) Jeremy: I think that you might be a bit too ambitious here. While I cannot speak to the anticipated execute time for a 20 atom mp2/6-31g* run, I can comment some on the required disk space: You are contemplating calculations involving on the order of 200 basis functions. The disk needed for just the mp2 energy is on the order of 200^3=8e6 64 bit words; or about 64 Mbytes. If you wish to do geometry optimizations, more disk will be needed, and if you need frequencies at this level (heaven forbid!) you may need on the order of 10 Gbytes of disk. I am not sure that it is possible to put this much storage on a PC, but even if it were, the bandwidth for getting to/from the disk will be much too slow. My guess is that the i/o time will increase the computed time (probably measured in days) by a factor of 10 or so. The PC version of g92 is probably best suited as a teaching tool, at which it is very good, rather than as production research tool. Good luck... Steve Williams chemistry, asu, boone, nc, usa ------------------------------------------------------------------------- From: "Sophie CREUZET ((33-1-40-01-56-39))" I have no experience concerning Gaussian '92 on a 486, working mainly on Supercomputer, but concerning the level of theory, I would not recommand the use of MP2 with a 6-31G* basis set. If you want to use MP2 you should put polarization functions on all atoms (i.e. 6-31G** in this case) unless your results might be even worse than at the 6-31G level. From jeremy@med.su.oz.au Mon Sep 26 22:15:19 1994 Received: from blackburn.med.su.oz.au for jeremy@med.su.oz.au by www.ccl.net (8.6.9/930601.1506) id VAA27826; Mon, 26 Sep 1994 21:08:55 -0400 Received: (jeremy@localhost) by blackburn.med.su.oz.au (8.6.8.1/8.6.4) id LAA12864 Tue, 27 Sep 1994 11:08:51 +1000 From: Jeremy R Greenwood Message-Id: <199409270108.LAA12864@blackburn.med.su.oz.au> Subject: Pyridazinedione response summary To: chemistry@ccl.net Date: Tue, 27 Sep 1994 11:08:51 +1000 (EST) Cc: hughc@blackburn.med.su.oz.au (Hugh Capper) X-Mailer: ELM [version 2.4 PL23] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 12856 Summary of responses to my query regarding the modelling of the tautomeric structures of aminomethyl pyridazine-3,6-diones: > I've been using MOPAC, however I don't know if I can trust the results > (based on W.Fabian, J. Molecular Structure (Theochem), 200 (1990), 295). > Which semi-empirical technique is most likely to give me reasonably > accurate structures and energies for this ring system? If I have > to resort to ab initio calculations, what is the minimum level > of theory necessary to handle this ring? > > Secondly, the structures I am modelling are most likely zwitterionic > in physiological solution, as they contain an amino group in addition > to the acidic pyridazindione. Is there any point in modelling zwitterions > on MOPAC or Gaussian 92 in the gas phase and trying to extrapolate > to solution behaviour? Is there an appropriate (not too computationally > expensive) technique for simulating solution behaviour, reliable for > predicting zwitterionic and tautomeric structure, or does one have to > resort to coordinating water molecules etc.? > > Thirdly, the atomic charges returned by various semi-empirical methods > vary widely for these structures. I would like to be able to predict > qualitatively if not quantitatively, which proton(s) are the most acidic. > I realise that this may once again depend on aqueous behaviour rather > than gas-phase, and hence be rather a tall order. Any suggestions? Jeremy Greenwood ---------------------------------------------------------------------------- From: "Raymond L. Roskwitakski" Subject: Re: Modelling Pyridazindiones Greetings Jeremy: I have some references: A prior pKa Calculations and Hydration of Organic Anions William L. Jorgensen J Am Chem Soc 1989 vol 111 pp. 4190-4197 Theoretical Examination of the S_N2 Reaction Involving Chloride Ion and Methyl Chloride in the Gas Phase and Aqueous Solution Jayaraman Chandrasekhar, Scott Smith, William L. Jorgensen J Am Chem Soc 1985 vol 107 pp. 154-163 Optimized Intermolecular Potential Functions for liquid Alcohols William L. Jorgensen J. Phys. Chem. 1986 vol 90 pp. 1276-1284 Relative Partion Coefficients for Organic Solutes from Fluid Simulations William L. Jorgensen James M. Briggs J. Phys. Chem. 1990 vol 94 pp. 1683-1686 Cis-Trans Energy Difference for the Peptidee Bond in the gas and Aqueous Solution William L. Jorgensen and Jiali Gao J Am Chem Soc 1988 vol 110 pp. 4212-4216 I don't know where Jorgensen is now. You can ask for his address on CCL. Best to talk with him to see if your problem can be handled well by this method. It seems the best way to me, but then again I am not the in Monte Carlo expert. He and Gao are. Your problem is not trvial, since the energy curves of tatomers are shallow. Inclusion of solvent effects makes sense to me, the gas phase results will be confusing, because different computer methods will give different results because of the shallowness. --------------------------------------------------------------------------- From: jborn@triton.unm.edu I recently modeled a series of alpha hydroxy lactones and I could not get either AM1 or MM2 to produce the hydrogen bond that seemed to be important in these molecules. I ended up using PC-Model and the results looked correct and match what physical characteristics that I know of. I wonder if there is a hydrogen bond between one of the NH groups and the alpha carbonyl in your system? Sincerely J. Born jborn@triton.unm.edu --------------------------------------------------------------------------- From: "Raymond L. Roskwitakski" Subject: RE: Modelling Pyridazindiones Hi Jeremy: In regard to part of your question: [ I would like to be able to predict qualitatively if not quantitatively, which proton(s) are the most acidic. I realise that this may once again depend on aqueous behaviour rather than gas-phase, and hence be rather a tall order. Any suggestions? ] I suggest perturbative free energy calculations using monte carlo to simulate the solvent. William Jorgenson, and Jiali Gao have done work in this. I believe Jorgenson has a program for this called BOSS. Hope this helps. Raymond Roskwitalski, Jr. rlr@autarch.acsu.buffalo.edu -------------------------------------------------------------------------- From cramer@maroon.tc.umn.edu Fri Sep 2 23:22:32 1994 Jeremy, It always seems a bit self-serving to me to refer to one's own work, however in this case it does seem particularly germane. We have explored exactly the issues you are interested in for some heterocycles in JACS 115 (1993) 8810-8817. Although the paper is far more thorough, let me say in a nutshell that (1) semiempirical methods are lousy for gas-phase energy differences unless you get lucky, (2) remarkably high levels of ab initio theory seem to be required if you want accurate relative energies for tautomers with differently hybridized heteroatoms, and (3) solvation models that account only for electrostatics (most continuum models) miss lots of the specific interactions from the first shell that are important for polar heterocycles, but explicit-solvent models (e.g., simulations) are terribly expensive and have electrostatic terms of questionable utility. Our own approach was to use our SM2 solvation model to correct gas-phase relative free energies, and that seemed to work well. As for atomic charges, we have recently developed something we call Charge Model 1 (a "Class IV" charge model) that maps semiempirical Mulliken charges to new values that we have shown to be similar in quality to electrostatic potential fitted charges from correlated, large-basis densities. This paper is presently in press at the Journal of Computer-Aided Molecular Design. Several heterocycles were included in the test set (although not quite so complex as a pyridazindione) and none seemed to be particularly problematic. I would be happy to send you a preprint if you are interested. Finally, from a practical standpoint, you are probably more interested in doing calculations than reading about them. Both the SM2 solvation model and the CM1 charge model (which can use either AM1 or PM3 wavefunctions, incidentally) are available in our program AMSOL, which is distributed by QCPE essentially for the medium cost (it may be about $200 at this point, I'm afraid I don't recall--anyway, we don't get any money . . . ) The address etc. for QCPE is in the CCL archives. Or, you could send e-mail to qcpe@ucs.indiana.edu. Best regards, Chris -- Christopher J. Cramer University of Minnesota Department of Chemistry 207 Pleasant St. SE Minneapolis, MN 55455-0431 (612) 624-0859 cramer@maroon.tc.umn.edu ------------------------------------------------------------------------- From MARTIN@cmda.abbott.com Fri Sep 2 23:22:55 1994 Let me know the answers to your questions. If you haven't seen it, the article by George Shields in Journal of Computtational Chemistry deals with H-bonds which might offer some hope to you. I wish I knew the answers. Also there was an article by Cramer and/or Truhlar dealing with some sort of tautomerization. I think it was in JACS. I'll hunt the references now. @@@@@@@@@@@ Yvonne Martin, Senior Project Leader @ Computer Assisted Molecular Design Project @@@@@@@@ @ D-47E, AP9A-LL @ @ Abbott Laboratories @ @ One Abbott Park Road @@@@@@@@@@ Abbott Park, IL 60064 Phone: 708 937-5362 FAX: 708 937-2625 ---------------------------------------------------------------------------- From stan@sun1.chem.univ.gda.pl Fri Sep 2 22:15:19 1994 Hi Jeremy For simulating solution (water) you can use two programs MOPAC93 (extended version of MOPAC 6.0 including method COSMO (J.Chem.Soc.Perkin.Trans. 2, 799-805 (1993)) or AMSOL (extended version of AMPAC 3.0 program including SM1, SM2 and SM3 methods for estimation of solvation energy (only for water solution)). All this programs are available from QCPE Catalogue. If you want use ab initio calculations see papers Schlegel at al. JACS 104, 5347 (1982) about tautomerisation of 2-Pyridone and 4-Pyridone, or Cieplak at al. Int. J. Quantum Chem. 14, 65 (1987) Reference for semi-empirical calculations of 2- and 4-pyridones: Katritzky and Szafran, J. Mol. Struct. (TCHEOCHEM) Stanislaw Oldziej Faculty of Chemistry University of Gdansk POLAND e-mail:stan@sun1.chem.univ.gda.pl ----------------------------------------------------------------------------- From markm@portal.vpharm.com Fri Sep 2 20:54:48 1994 Dear Jeremy: Interesting problem. I'm surely no expert in this field, but based on limited experience you might wish to consider MP2/6-31G* optimizations. The system is pretty small, so I think this would be possible with gaussian-92 running on (say) an IBM550 or greater. You will need lots of disk space; 2 Gb should do. If the MP2 optimization is not affordable, the other approach (which makes sense to do anyway) is a series of opts with increasing basis sets to see how your results change from one to the next. Start with 6-31G* then 6-31+G* then 6-31++G* ... etc. You might also look at DH basis sets. The energy differences between tautomers is often quite small and correlation seems like something you definitely want to include, so follow the HF opts with some kind of correlated single points. You also could follow the MP2/6-31G* opts (if you do them) with higher basis-set correlated single points calcs. I will avoid the whole religious issue of which correlation method would be best for your system. I have no idea whether density functional is appropriate for this kind of system. Semi-empirical is a complete waste of time, IMHO. Finally, I gather the formula is C4H4N2O2. What is your experimental evidence for the hydroxy form? Is it the diol or the mono-ol? What is the pKa? Pyridones (sometimes called 2-hydroxy pyridines) can be drawn as hydroxy compounds, but in fact they exist as the lactams. The proton is on the nitrogen. So if you actually have the alcohol form, then somehow this system's extra nitrogen atom has switched the tautomeric preference. That would be interesting to understand. (I freely confess I don't understand tautomerization very well.) J: Actually the formula of the structures I'm dealing with is J: C5H7N3O2, containing an aminomethyl substitution. The literature J: sugests that in gas, liquid, and solid phases, for pyridazine-3,6-diones, J: the oxo/hydroxy (mono-ol, mono-one) tautomers are preferred. This J: is indeed interesting, and somewhat different from other hydroxy- J: nitrogen heterocycles. Pyridazine-3,6-dione itself is strongly acidic, J: pKa = 1.00 I would love to learn more about your system, and of course look forward to your summary! / Mark (markm@vpharm.com) ----------------------------------------------------------------------- From inoue@greencross.co.jp Fri Sep 2 21:13:36 1994 Dear Dr.Jeremy Greenwood, >Which semi-empirical technique is most likely to give me reasonably >accurate structures and energies for this ring system? I think PM3 gives better 'total energies' than AM1 or MNDO, however, do not know about the structures. >Is there an appropriate (not too computationally expensive) technique >for simulating solution behaviour, reliable for predicting zwitterionic >and tautomeric structure? If you are using MOPAC93 please add EPS=78.4 to the keyword line. >I would like to be able to predict qualitatively if not quantitatively, >which proton(s) are the most acidic. How about removing one H-atom from the molecule, adding CHARGE=-1 and EPS=78.4 to the keyword line. If you repeat this and evaluate each 'heat of formation' data, you could say which state is stable, it means the most acidic proton. Hope this help. ____/ ___/ ___/ Yoshihisa INOUE (^_^) the Green Cross Corp. / / / 2-25-1 Shodai-Ohtani,Hirakata,Osaka 573 JAPAN / _ / / / tel: +81-720-56-9328 / / / / fax: +81-720-68-9597 _____/ _____/_____/ E-mail: inoue@greencross.co.jp ----------------------------------------------------------------------------- From newhoir@duc.auburn.edu Sat Sep 3 00:36:35 1994 To answer one point of many in your question: AMSOL is capable of modelling solution behavior. It is even possible to compute zwitterions in the gas phase: I just saw a poster on the topic at the Washington ACS meeting for glycine. (intramolecular proton transfer). The authors published this in JACSS, at the end of '93, I believe. There have been many discussions on computing partial charges on this mailer: you should check the archives! Good luck! Irene Newhouse ----------------------------------------------------------------------------- From tjm@chem.ucla.edu Tue Sep 27 00:00:43 1994 Received: from uclachem.chem.ucla.edu for tjm@chem.ucla.edu by www.ccl.net (8.6.9/930601.1506) id XAA29462; Mon, 26 Sep 1994 23:05:07 -0400 From: Received: from theory.chem.ucla.edu by uclachem.chem.ucla.edu (Sendmail 5.61/1.05) id AA23744; Mon, 26 Sep 94 20:00:41 -0700 Message-Id: <9409270300.AA23744@uclachem.chem.ucla.edu> Received: by theory.chem.ucla.edu (5.51/celerity1.2) id AA02294; Mon, 26 Sep 94 19:55:39 PDT Subject: FORTRAN on PC To: chemistry@ccl.net Date: Mon, 26 Sep 94 19:55:38 PDT X-Mailer: ELM [version 2.3 PL2] Does anyone have experience with the various FORTRAN compilers for the PC, expecially with regards to Pentium performance? I am considering running some quantum codes on a Pentium 90 and wonder if there is much difference between the CPU times for such codes compiled with different vendor compilers. I know of the Microway, Lahey, Watcom, and Microsoft Powerstation compilers, but do not know anything about their relative performances on ab initio style codes. Am not too concerned with user-friendliness, but will appreciate such comments also. -Todd Martinez (tjm@theory.chem.ucla.edu) From cornell@cgl.ucsf.EDU Tue Sep 27 01:00:51 1994 Received: from socrates.ucsf.EDU for cornell@cgl.ucsf.EDU by www.ccl.net (8.6.9/930601.1506) id AAA00263; Tue, 27 Sep 1994 00:38:23 -0400 Received: (from cornell@localhost) by socrates.ucsf.EDU (8.6.9/GSC4.24) id VAA18473 for chemistry@ccl.net; Mon, 26 Sep 1994 21:38:21 -0700 Date: Mon, 26 Sep 1994 21:38:21 -0700 Message-Id: <199409270438.VAA18473@socrates.ucsf.EDU> From: cornell@cgl.ucsf.edu (Wendy Cornell) To: chemistry@ccl.net Subject: Charges A number of posters have asked about atom-centered charges. I discuss here their derivation for use in molecular mechanical calculations. The philosophy of the Kollman group (AMBER) has been that the accurate representation of electrostatic interactions is crucial for a force field intended for application to biological molecules. "We note that the choice of a particular force field should depend on the system properties one is interested in. Some applications require more refined force fields than others. Moreover, there should be a balance between the levels of accuracy or refinement of different parts of a molecular model. Otherwise the computing effort put into a very detailed and accurate part of the calculations may easily be wasted due to the distorting effect of the cruder parts of the model." (van Gunsteren, W.F and Berendsen, H.J.C., Angew. Chem. Int. Ed. Engl. 29, 992 (1990) -- an incredibly lucid and succinct review of MD applications to chemistry). In other words, a force field which has a complicated potential form for representing bonds and angles and is very precise in terms of reproducing geometries and vibrational frequencies will not accurately model complex intermolecular interactions if the charge model is not also of high quality. Piotr Cieplak and I have derived charges for the new "AMBER" force field. We have already published two papers describing the new AMBER approach to calculating ESP (electrostatic potential) fit charges. The new charges are called RESP charges, for Restrained ESP fit. This method was developed by Chris Bayly who was a postdoc in our group. The 2 references for these papers are: 1. BAYLY CI; CIEPLAK P; CORNELL WD; KOLLMAN PA. A WELL-BEHAVED ELECTROSTATIC POTENTIAL BASED METHOD USING CHARGE RESTRAINTS FOR DERIVING ATOMIC CHARGES - THE RESP MODEL. JOURNAL OF PHYSICAL CHEMISTRY, 1993 OCT 7, V97 N40:10269-10280. 2. CORNELL WD; CIEPLAK P; BAYLY CI; KOLLMAN PA. APPLICATION OF RESP CHARGES TO CALCULATE CONFORMATIONAL ENERGIES, HYDROGEN BOND ENERGIES, AND FREE ENERGIES OF SOLVATION. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY, 1993 OCT 20, V115 N21:9620-9631. Earlier, I posted a summary of our new charge model and I would be happy to mail a copy to any interested parties who missed that posting. I believe the original work in the area of ESP fitted charges was carried out by Frank Momany: Momany, F., J. Phys. Chem., 82, 592 (1978). This work was later extended by others: Cox, S.R. and Williams, D.E., J. Comp. Chem., 2, 304 (1981) Singh, U.C. and Kollman, P.A., J. Comp. Chem., 5, 129 (1984) (Gaussian80/UCSF) Chirlian, L.E. and Francl, M.M., J. Comp. Chem., 8, 894 (1987) (CHELPG) Williams, D.E., Biopolymers, 29, 1367 (1990) Bessler, B.H., Merz, K.M., Jr. and Kollman, P.A., J. Comp. Chem., 11, 431 (1990). (MOPAC) Gaussian 92 includes an option for calculating "Merz-Singh-Kollman" CHELP, and CHELPG style charges. The basic idea with electrostatic potential fit charges is that a least squares fitting algorithm is used to derive a set of atom-centered point charges which best reproduce the electrostatic potential of the molecule. In the AMBER charge fitting programs, the potential is evaluated at a large number of points defined by 4 shells of surfaces at 1.4, 1.6, 1.8, and 2.0 times the VDW radii. These distances have been shown to be appropriate for deriving charges which reproduce typical intermolecular interactions (energies and distances). The dipole moment of the molecule is well reproduced. Other programs have embedded the molecule in a cubic grid of points to evaluate the potential. We believe that assigning the points along the contours of the molecule provides a better sampling of the esp around each atom. The value of the electrostatic potential at each grid point is calculated from the quantum mechanical wavefunction. The charges derived using this procedure are basis set dependent. For example, the Weiner et al force field (AMBER) employs STO-3G based charges, whereas the new Cornell et al force field (AMBER) uses charges derived using the 6-31G* basis set. The 6-31G* basis set is bigger and, for the most part, "better." Because quantum mechanics calculations scale as the number of basis functions to about the 2.7 power (HF as implemented in G92), the bigger 6-31G* basis set was prohibitively large for use in developing the last force field. The 6-31G* basis set tends to result in dipole moments which are 10-20% larger than gas phase. This behavior is desirable for deriving charges to be used for condensed phase simulations within an effective two-body additive model, where polarization is being represented implicitly. In other words a molecule is expected to be more polarized in condensed phase vs. gas phase due to many body interactions, so we "pre-polarize" the charges. A study by St-Amant, Cornell, Halgren, and Kollman (submitted) calculated DFT charges for a number of small molecules and found them to be smaller than HF/6-31G* derived ones. DFT charges for methanol did not reproduce the relative free energy of solvation of methanol. Such charges may be more appropriate for use with a non-additive model. (I should note that the DFT model reproduced the gas phase dipole moments very well.) ESP fit charges have many advantages. They reproduce interaction energies well. They can be calculated in a straightforward fashion. They have been shown to perform well at reproducing conformational energies. The second paper listed above (Cornell et al, JACS) provides much of the validation of our new charge model. A study by Howard, Cieplak, and Kollman (J Comp Chem, in press) showed how ESP and RESP charges performed quite well at modelling the conformational energies of a series of 1,3-dioxanes. Also, a more thorough study of the performance of RESP charges at calculating small molecule conformational energies is currently underway in our group. It should be noted that Mulliken charges do NOT reproduce the electrostatic potential of a molecule very well. Mulliken charges are calculated by determining the electron population of each atom as defined by the basis functions. When the density is associated with the square of a single basis function, that density is assigned to the atom associated with that basis function. Similarly, if the density is associated with 2 basis functions which are on a common atom, the density is assigned to that atom. The ambiguity arises when the density is associated with 2 basis functions lying on different atoms. In that case the density is partitioned equally onto each atom. Another charge model is that of Gasteiger-Marsili. (Gasteiger and Marsili, Tet. Lett., 36, 3219 (1980)). This approach involves the partial equalization of electronegativity between bonded atoms. Finally, it's a little dangerous to look too closely at charges. Sometimes ones that look "funny" or "too big" actually perform quite well at reproducing the desired properties. ----------------------------------------------------------------------- Wendy D. Cornell Graduate Group in Biophysics Box 0446 (415) 476-2597 (phone) Department of Parmaceutical Chemistry (415) 476-0688 (fax) University of California, S.F. cornell@cgl.ucsf.edu San Francisco, CA 94143-0446 USA From qftjesus@usc.es Tue Sep 27 06:00:52 1994 Received: from uscmail.usc.es for qftjesus@usc.es by www.ccl.net (8.6.9/930601.1506) id FAA02588; Tue, 27 Sep 1994 05:28:34 -0400 Received: by uscmail.usc.es (4.1/1.34) id AA01763; Tue, 27 Sep 94 10:29:42 +0100 Date: Tue, 27 Sep 94 10:29:42 +0100 From: qftjesus@usc.es (Jesus Otero Rodriguez) Message-Id: <9409270929.AA01763@uscmail.usc.es> To: chemistry@ccl.net Subject: NBO manual Dear netters: I would like to get the NBO manual because it is not included in the Gaussian 92 manual (or at least I do not find it). Thanks e-mail: qftjesus@usc.es From noy@tci005.uibk.ac.at Tue Sep 27 11:01:08 1994 Received: from uibk.ac.at for noy@tci005.uibk.ac.at by www.ccl.net (8.6.9/930601.1506) id KAA07515; Tue, 27 Sep 1994 10:28:06 -0400 Received: from tci005.uibk.ac.at by uibk.ac.at with SMTP id AA08633 (5.65c/IDA-1.4.4 for ); Tue, 27 Sep 1994 15:27:55 +0100 Received: by tci005.uibk.ac.at (AIX 3.2/UCB 5.64/CFIBK-2e.AIX) id AA36956; Tue, 27 Sep 1994 15:27:56 +0100 From: noy@tci005.uibk.ac.at (Noy) Message-Id: <9409271427.AA36956@tci005.uibk.ac.at> Subject: NBO To: chemistry@ccl.net Date: Tue, 27 Sep 1994 15:27:55 +0100 (NFT) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 932 Hallo cyberchemists, A review by A.E. Reed on Chem.Rev. 88:899-926 (1988) is also a good place to go for obtaining references therein about NBO. Reed, A.E., Curtiss, L.A. and Weinhold, F. (1988) Inter molecular Interactions from a Natural Bond Orbital, Donor- Acceptor Viewpoint. Chem.Rev. 88:899-926. best wishes, Noy ---------------------------------------------------------------------------- Teerakiat Kerdcharoen (NOY) Institute of General and Inorganic and Theoretical Chemistry Innrain 52a, A-6020 Innsbruck AUSTRIA e-mail: noy@tci2.uibk.ac.at, noy@tci.uibk.ac.at, c72454@cx.uibk.ac.at : noy@atc.atccu.chula.ac.th, noy@atc2.atccu.chula.ac.th ( Bangkok ) Research : Molecular Dynamics simulations : Computer Aided Molecular/Material Designs ----------------------------------------------------------------------------- *** I have no past and no future. I just have today. From kottalam@rhea.cray.com Tue Sep 27 11:07:38 1994 Received: from timbuk.cray.com for kottalam@rhea.cray.com by www.ccl.net (8.6.9/930601.1506) id KAA07735; Tue, 27 Sep 1994 10:41:25 -0400 Received: from wake.cray.com (kottalam@wake.cray.com [128.162.172.2]) by timbuk.cray.com (8.6.9/8.6.9L) with SMTP id JAA09146 for ; Tue, 27 Sep 1994 09:41:25 -0500 Received: by wake.cray.com (4.1/CRI-5.14) id AA01050; Tue, 27 Sep 94 09:46:33 CDT Date: Tue, 27 Sep 94 09:46:33 CDT From: kottalam@rhea.cray.com (Jeyapandian Kottalam) Message-Id: <9409271446.AA01050@wake.cray.com> To: chemistry@ccl.net Subject: CCL: New Amber speed record + PC timing Cc: kottalam@rhea.cray.com No need to accept personal bias. > From: rbw@msc.edu (Richard Walsh) > > > From: ross@cgl.ucsf.edu (Bill Ross > > > > From time to time in the past I have sent my latest benchmarks > > for Amber (molecular mechanics) when a new machine breaks the > > speed record. > > The SGI TFP performance numbers are impressive, but they should be > presented here with as high a regard for completeness as for enthusiasm > > Accepting the possibility of some personal bias :-)-- We are optimizing Amber4.1 for Cray C90. For the first set of three jobs, we get C90 31.9 32.2 32.3 Benchmarks involving 100 energy evaluations do not provide fair comparisons between vector and scalar machines. Prosessing the molecular topology and setting up the list of bonds, ect. is done only once. This is scalar code. The Cray does the minimization and dynamics fast after setting up the job in scalar mode. Realistic user jobs would involve several thousands of energy evaluations. Kottalam Phone: (USA) 612 683 3622 Senior Computational Chemist Cray Research, Inc. From qftramos@usc.es Tue Sep 27 14:01:05 1994 Received: from uscmail.usc.es for qftramos@usc.es by www.ccl.net (8.6.9/930601.1506) id NAA13688; Tue, 27 Sep 1994 13:47:38 -0400 Received: by uscmail.usc.es (4.1/1.34) id AA14862; Tue, 27 Sep 94 18:48:44 +0100 Date: Tue, 27 Sep 94 18:48:44 +0100 From: qftramos@usc.es (Antonio Fernandez Ramos) Message-Id: <9409271748.AA14862@uscmail.usc.es> To: chemistry@ccl.net Subject: IRC with GAUSSIAN92 Dear netters: I've been trying IRC calculations with G92. There is no problem from TS to reactives or products, but I'd like to be able to follow the lowest frequency normal mode from reactives or products to the opposite side where the TS is. Is there any option(s) with G92 to follow a normal mode this way? Thanks Antonio F. Ramos E-mail qftramos@usc.es From ross@cgl.ucsf.EDU Tue Sep 27 17:10:52 1994 Received: from socrates.ucsf.EDU for ross@cgl.ucsf.EDU by www.ccl.net (8.6.9/930601.1506) id QAA18233; Tue, 27 Sep 1994 16:58:46 -0400 Received: (from ross@localhost) by socrates.ucsf.EDU (8.6.9/GSC4.24) id NAA10923; Tue, 27 Sep 1994 13:58:39 -0700 Date: Tue, 27 Sep 1994 13:58:39 -0700 Message-Id: <199409272058.NAA10923@socrates.ucsf.EDU> From: ross@cgl.ucsf.edu (Bill Ross ) To: chemistry@ccl.net Subject: parallelization & benchmarks Cc: ross@cgl.ucsf.EDU I think my 'speed record' could use some more analysis to put it into context in the real world where everyday jobs are run. I also describe some improvements in the next version of the benchmarks. I believe that Cray's relative lack of emphasis on shared memory parallelization is not a major handicap: have you ever seen a Cray with fewer jobs than processors? Given the necessary inefficiency of running parallel, a job-loaded machine is better off running serially, as long as it can fit Ncpu jobs into memory at once (Crays don't use virtual memory so must fit the whole job into memory). But if there's even one job that's big enough to keep the loaded jobs < Ncpu, then parallel begins to make sense. So on the typical loaded shared memory machine, one would have to examine the memory mix. I notice that an Amber job w/ 10K atoms takes 0.2% of the memory on a 1GB Convex, so in the real world I would put shared memory parallel of Amber on a Cray down as a frill unless there are more idle Crays than I expect, although I have heard of one large Cray shop where it might be useful. An SGI is somewhat less likely to be loaded with multiple jobs, in my experience, which in combination with cheaper processors provides more of an incentive to parallelize, and large quantum calculations would probably also be worth parallelizing on Crays. By the same token, although we are working on message-passing parallelism, we are not expecting to use it on our HP cluster, since there is usually at least 1 job on each machine and thus parallelization would reduce throughput (we will, however, use it on Paragon, SP2 and T3D machines). We'd need a sophisticated batch scheduler to balance onto & off of machines so that a process could 'go parallel' when a machine became idle. In fact, if anyone knows about parallel job schedulers, I would be quite interested in hearing from them. Presumably this load-balancing issue would apply to shared memory scheduling, though there the operating system kernel might handle it. Richard Walsh has just informed me that the 'Linda' people claim to have taken care of the load balancing (phone 203 777-7442 for those who want to investigate). Barry Bolding and Jeyapandian Kottalam at Cray and Steve Chin at IBM have remarked that the shortness of the benchmarks (100 steps each) cause a greater percentage of time to be spent in scalar setup of the job as speeds increase. This is quite a valid point: free energy calculation on DNA in vacuum --- time (sec) --- startup total % startup HP 735 3 174 1.7 Cray C90 2 66 3.0 The free energy perturbation uses about twice the runtime as plain dynamics because of multiple energy evaluations, but it has to set up more as well. Clearly the percentage of setup time is worth considering further; my immediate reaction has been to increase the number of steps to 500. Thomas Huber of Ludwig Maximilian Universitaet, Muenchen has suggested subtracting the time for a 100 step run from a 200 step run, and perhaps this is what I'll do instead. I don't see any need for thousands of steps unless it proves necessary to average out fluctuations in the rest of the computer system - which may be required for the message- passing architectures, especially workstation clusters (the results I have reported so far are rather reproducible). Much more important in the general case is to vary the cutoff distance. At this point I'd like to thank again the growing list of people who have contributed by discussion or remark to my formulation and interpretation of the benchmarks. In approximate temporal order: George Seibel (now at Smith Kline), Dave Case (Scripps), Steve DeBolt (Scripps), Steven Chin (IBM), Thomas Huber (Ludwig Maximilian Universitaet, Muenchen), Richard Walsh (Minnesota Supercomputer Center, Inc.), Barry Bolding (Cray) and Jeyapandian Kottalam (Cray). Bill Ross From cornell@cgl.ucsf.EDU Tue Sep 27 23:10:41 1994 Received: from socrates.ucsf.EDU for cornell@cgl.ucsf.EDU by www.ccl.net (8.6.9/930601.1506) id WAA23110; Tue, 27 Sep 1994 22:44:21 -0400 Received: (from cornell@localhost) by socrates.ucsf.EDU (8.6.9/GSC4.24) id TAA15350 for chemistry@ccl.net; Tue, 27 Sep 1994 19:44:21 -0700 Date: Tue, 27 Sep 1994 19:44:21 -0700 Message-Id: <199409280244.TAA15350@socrates.ucsf.EDU> From: cornell@cgl.ucsf.edu (Wendy Cornell) To: chemistry@ccl.net Subject: charges I've gotten some encouraging feedback regarding the usefulness of my previous post so I am now going to post a description of our (Kollman group) current charge model (RESP) and also a a discussion of some of the pitfalls of esp-fit charges. MULTIPLE CONFORMATION RESP MODEL Our current charge model is the RESP (restrained ESP) model. This model evolved from work by Chris Bayly which showed that charges on buried atoms (such as alkyl carbons) were not well determined by the electrostatic potential points. Such buried charges often assumed large values during the fitting process and the values of these charges showed great conformational variability. The basic idea of the RESP model is that restraints are added to charges on non-hydrogen atoms during the charge fit. The charges are restrained to an "optimal" value of zero. The restraints are hyperbolic in nature, so approximately the same amount of force is felt by charges of all magnitudes. An earlier model employed harmonic restraints, but they reduced the values of the heteroatom charges too drastically since those values (typically +/- 0.6 or higher) fell in the steep part of the function. The details of the derivation of this method are given in the Bayly et al JPC paper. The RESP method involves a two-stage approach where charges on atoms such as methyl hydrogens are not forced to be equivalent until the second stage. At that point they are refit while charges on the other atoms are constrained to their values from stage one. Forcing methyl hydrogens to have equivalent charges during the first stage can adversely affect the values of the heteroatom charges, because such hydrogens are not equivalent in a static conformation. In the standard ESP model, methyl hydrogen charges were typically averaged after the fit, but this averaging often changed the value of the dipole moment as well as the fit to the potential. One problem with electrostatic potential fit charges in general is that they reproduce the molecular potential and the dipole moment very well *for the conformation of the molecule employed in the fit*. However, when those charges are applied to other conformations, the agreement is not as good. A solution to this problem was proposed by Chris Reynolds: REYNOLDS CA; ESSEX JW; RICHARDS WG. ATOMIC CHARGES FOR VARIABLE MOLECULAR CONFORMATIONS. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY, 1992 NOV 4, V114 N23:9075-9079. The solution is simply to use multiple conformations of a molecule in the charge fitting process. In fitting the amino acid charges for our new force field, we used 2 conformations for each amino acid -- the first conformation had the backbone in an extended conformation and the second had it in an alpha-helical conformation. Each of those 2 conformations then had different values for chi1, chi2, etc. I should repeat that for our new force field we are using 6-31G* fit charges because that basis set provides the enhanced polarity necessary for carrying out balanced (with respect to the water model) condensed phase simulations within an additive model (no polarization). Charges derived using a higher level of theory (either in terms of a bigger basis set or through the inclusion of correlation) won't necessarily be better for such applications if they do not result in dipole moments which are enhanced over the gas phase values. We have found that multiple conformation RESP fitting results in fairly robust charges. Software for carrying out these fits will be available with the next release of AMBER. The new software (C. Bayly) also allows for specifying additional lagrange constraints so that (e.g.) blocking groups can be forced to have net neutral charges and molecules can be spliced together more algorithmically. PITFALLS OF esp-FIT CHARGES In general, standard esp-fit charges perform well at reproducing desired properties such as DNA base pair, NMA dimer, and methanol-water interaction energies. This is not always the case, however. For example, Yax Sun in our group calculated ESP charges for a spherand for the purpose of carrying out free energy perturbation studies to compare the binding of Li+ and Na+ to the spherand. The spherand looked like this: /\ { | |_ } \/ 6 | O \ CH3 Standard ESP charges were calculated from a 3-unit, non-cyclic, methyl-blocked analog. The charges were then taken from the central residue. The standard ESP fit charges underestimated the interaction energies between the spherand and the ions. As the charge on the oxygen was on the low side, this was thought to be the source of the error. Sun and Kollman then refit the charges, this time with electrostatic potential points within 5 A of the oxygen weighted more heavily than those around the rest of the molecule. The resulting charges resulted in a relative free energy of binding in much better agreement with experiment. There also cases where polarization or lone pairs are required to reproduce interaction energies. Electrostatic potential fit charges also do not always result in good conformational energies. For example, multiple conformation (C5/aR) RESP charges calculated for glycine and alanine dipeptides do not result in conformational energies which are in good agreement with the quantum mechanically calculated values. These charges were derived using the 6-31G* basis set and were applied with a 1-4 electrostatic scale factor of 1/1.2 (Cornell et al JACS). The reason for this performance is unclear. It is possible that the 6-31G* charges overstabilize the C7 (7-membered H-bonded ring) conformations. When these charges are scaled back to gas-phase-like values (q*0.88), the conformational energies show good agreement with the QM data. My statement in the previous post that ESP/RESP charges resulted in good conformational energies referred to the 6-31G* based model with the 1-4 elect. scale factor of 1/1.2 I've given 2 examples where esp-fit charges were not able to be applied in a straighforward fashion. Overall, however, we have found these charge models (ESP and RESP) to be quite useful for modelling biomolecular systems. The alternatives usually involve either (1) models such as Mulliken charges which do not necessarily reproduce the molecular electrostatic potential or (2) empirically derived charges such as those fit to reproduce interaction energies and distances (CHARMM) or liquid properties (OPLS). Because electrostatic potential fit charges can be calculated fairly easily, they allow the force field to be extended to other molecules. Overall we find them to be a very useful and relatively general model. ----------------------------------------------------------------------- Wendy D. Cornell Graduate Group in Biophysics Box 0446 (415) 476-2597 (phone) Department of Parmaceutical Chemistry (415) 476-0688 (fax) University of California, S.F. cornell@cgl.ucsf.edu San Francisco, CA 94143-0446 USA