From vis@gensia.com Tue Mar 5 01:32:24 1996 Received: from gtc05f for vis@gensia.com by www.ccl.net (8.7.1/950822.1) id AAA21417; Tue, 5 Mar 1996 00:38:51 -0500 (EST) Received: from genny (genny.res.gensia.com) by gtc05f (5.x/SMI-SVR4) id AA29751; Mon, 4 Mar 1996 21:32:47 -0800 Received: by genny (940816.SGI.8.6.9/930416.SGI.AUTO) id FAA10929; Tue, 5 Mar 1996 05:05:52 GMT Date: Mon, 4 Mar 1996 21:05:51 +48000 From: Viswanadhan To: CHEMISTRY@www.ccl.net Subject: Responses to my query and a question Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear Netters, In response to my question, I received some responses summarized below. The question was: ------------------- We have a fairly rigid cyclic molecule (C) which can exist in two conformations (puckers), C1 and C2. Its acyclic analog (A) can also exist in different conformations, but they are not separated by high barriers. C1 and some conformers of A appear to have good interactions with a protein, but not C2. In terms of binding geometry/energies C1 is better than A but A is much better than C2. C1 and C2 are separated by a high barrier, but they are isoenergetic (internal energy from QM calculations). TCP (free energy perturbation) calculations for the mutation C1 -> A (or C2 -> A) will not reveal much because of the high barrier between C1 and C2 (making interconversion almost impossible in a reasonable time). Do you think it is possible to qualitatively assess the binding of C vs. A ? ----------------- The question I have for today is what is the best and most efficient method to calculate the transion state barrier (and free energy profile) between two puckers for a cyclic molecule. I am aware of LST in Gaussian. I need a method that is accurate and works well for cyclic structures even if it takes a good deal of computer time. I will summarize the responses. ----------------- This is what I think about my previous query and responses. While the answers are interesting, it seems to me that if I can correctly compute the barrier and the torsion profile, I may guestimate the relative frequency (or probability) of interconversion (C1 to C2) relative to the acyclic molecule with a smaller barrier. Computing interaction energies in the complex and in the solvent for each molecule (one conformation in the complex as the start, and both conformers in the solvent) would enable the estimation of relative difference in binding enthalpy. If C is much better than A, and the torsional barrier is not too high, I would expect that C should be better than A as an inhibitor, because A would lose out on the entropic argument as well. While there is some subjectivity in this process, it will lead to better design ideas such as how the binding conformation can be further restricted, and what changes in the structure would lower the barrier etc. Monte Carlo methods are promising however, they are not easily applied to endocyclic torsions when dealing with protein-ligand complexes. The work of Dr. Still and coworkers is useful as a start in this respect. Vellarkad Viswanadhan vis@gensia.com ------------------------ Responses and Comments. Regarding your C1, C2, A problem: if you could compute the difference in chemical potential (free energy) between C1 and C2, you'd be all set. It might be possible to compute the work of forcing this transition, via a free energy calculation. Mike Gilson Computing the barrier would certainly help. However, it is not easy to accomplish the free energy calculation via umbrella sampling etc, though in principle this should work. From: gilson@indigo14.carb.nist.gov (Michael K. Gilson) To: vis@gensia.com, gilson@indigo14.carb.nist.gov Message-Id: <9602292200.AA04482@indigo14.carb.nist.gov> Content-Transfer-Encoding: 7BIT Status: RO X-Status: Dear Viswanadhan, If you know the equilibrium constants for binding of C1 and C2 to the receptor, and also that for the interconversion of C1 and C2, then you can use algebra to compute the "overall" or "apparent" equilibrium for binding of C to the receptor. Does this sound right? Mike To: Viswanadhan Status: RO X-Status: It is theoretically possible to do what you want directly with Monte Carlo calculations rather than FEP theory. For example, you can sample conformations of C1 and C2 in the bound form and determine the ratio existing at equilibrium and do the same without the protein. This can be done with continuum solvation models such as Clark-Still's. This kind of work is going on here at Mt. Sinai for peptides in water but could be extended to include protein. For the other problem of the binding of A versus C you could do a Monte Carlo calculation sampling AP versus A + P and another sampling CP versus C + P. You can use internal coordinates in the sampling and restrict the range of values sampled to include only those which exist in the free and bound forms of these molecules. This improves the sampling efficiency of Monte Carlo greatly over blind approaches and local sampling methods such as force bias. I'm not aware of anyone having tried exactly what I've described yet but it might work. Sincerely, Dave Garmer Physiology & Biophysics Dept. Mt. Sinai Medical School New York, NY 10029 We have had positive experience with GB/SA model in water for small molecules. However, protein-ligand calculations have strong limitations with continuum models, especially if the ligand is at the solvent interface and it is charged. If we let the side chains be flexible (some of which are charged), you will see big movements and large rms deviations. Our own work on a database of twenty proteins convinced us that improvements are necessary in the methods to apply MC methods+continuum models successfully (for the bound protein). However, your approach should work with good params and models. -------- Also, Well, you expect the polar and charged groups in solution to be surrounded by ions in solution. Since we do not put those, the sidechains tend to make internal salt bridges which are unrealistic. This will also be at the expense of right geometry for ligand and protein sidechains (interactions) since these are at the surface. With regard to the internal sites, if they are far away from solvent, why do you think solvent interactions would play an important role in binding? It is a quite different matter if you are modeling the whole flap opening or something like that. From: Dave Garmer To: Viswanadham Status: RO X-Status: Dear Viswanadham, To answer the easy question first, there is evidence of ordered water molecules around the retinal in rhodopsin and the modeling work also suggests that there are enough unsatisfied hydrogen bonds to generate a water network around GPCR ligands. I would guess that an exposed site would have less ordering and so better described as a polarizable continuum. Regarding crystallographic protein surface details, the requirements for crystallization usually mean that the local ion concentration is high compared to an organism. Also, crystallization is converting molecules which exist in a liquid environment into a solid, thus probably eliminating some inherent flexibility. Also, crystallographers try to assign what they think is the best single structure and very seldom are able to identify what might be minority structures for small parts of the protein. Theoretical studies have indicated that there probably are minority structures often existing even for internal packed sidechains. Therefore, I would trust crystallography only to suggest particular accessible structures for flexible units. I would also think that a decent continuum model like Clark-Stills probably can more often than not give you a sample of some of the thermally accessible structures. They may not be the same as the ones being sampled in a crystal structure refinement. Dave Garmer From: "E. Lewars" Message-Id: <199602281619.LAA26845@alchemy.chem.utoronto.ca> To: vis@gensia.com Subject: SUBSTRATE CONFORMERS/PROTEIN Status: RO X-Status: It seems to me you should first try to find the global minimum. If this is also the best-binding conformer, then the barriers to conformer interconversion are not relevant to the problem (I think). If the global minimum is _not_ the the conformer that binds strongest, then you might have to write out a scheme involving all the species and analyze it as a problem in (mainly) kinetics. Sorry, but I haven't looked at this in detail. Errol Lewars ==== This is rigid molecule (C) with only two minima which are equivalent (internal energy-wise). Only one is the binding conf. The question is C vs. A. I am not sure what you precisely mean by 'problem in kinetics'. Is there any similar problem that is analysed? From: "E. Lewars" Message-Id: <199602281821.NAA11848@alchemy.chem.utoronto.ca> To: vis@gensia.com Subject: CONFORMERS Status: RO X-Status: Hello, I thought that the acyclic molecule A was in equilibrium with C; evidently this is not so, i.e. A is a quite separate species. If C1 and C2 are in equilibrium and only C1 binds to the protein, then if enough protein is present all the C1 will get tied up and then C2 will (slowly, if the C2--> C1 barrier is high) drain away to C1.protein. To see if this happens on a useful timescale, you have to calc the C1/C2 barrier and estimate the half-life of C2. Ab initio calcs give more reliable results for barriers than do semiempirical calcs. Good luck Errol Lewars === From jaouad@goliat.ugr.es Tue Mar 5 04:32:27 1996 Received: from goliat.ugr.es for jaouad@goliat.ugr.es by www.ccl.net (8.7.1/950822.1) id DAA23457; Tue, 5 Mar 1996 03:56:36 -0500 (EST) Received: (from jaouad@localhost) by goliat.ugr.es (8.6.10/8.6.12) id JAA29623; Tue, 5 Mar 1996 09:57:39 GMT Date: Tue, 5 Mar 1996 09:57:38 +0000 (WET) From: Jauoad El Bahraoui To: chemistry@www.ccl.net Subject: Mobius Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear Netters: I'm trying to perform some calculations of Mobius cyclic molecules - can sombody point me to reference of previous therotical calculation? - can somebody say me how to construct a z-matriz for this mentioned species? thanks in advance. Jaouad El Bahraoui Email: jaouad@goliat.ugr.es From p.mencarelli@caspur.it Tue Mar 5 04:41:11 1996 Received: from axcasp.caspur.it for p.mencarelli@caspur.it by www.ccl.net (8.7.1/950822.1) id EAA23530; Tue, 5 Mar 1996 04:14:10 -0500 (EST) Date: Tue, 5 Mar 1996 04:14:10 -0500 (EST) Message-Id: <199603050914.EAA23530@www.ccl.net> Received: from aldebaran.sci.uniroma1.it by axcasp.caspur.it with SMTP; Tue, 5 Mar 1996 10:13:33 GMT X-Sender: mencarel@axcasp.caspur.it X-Mailer: Windows Eudora Light Version 1.5.2 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: chemistry@www.ccl.net From: "Dr. Paolo Mencarelli" Subject: Summary:Interaction energy in host-guest complexes Dear CClers, Some time ago I posted a question about the interaction energy in host-guest complexes. Here are the answers I got: ============================================================================ ========= My original question: Hi all, I am an organic chemist interested in host-guest chemistry and I have a question on a computational aspect of this topic. The interaction energy (Delta E) of the host and the guest in a given complex may be defined as: Delta E = Esteric(complex) - Esteric(host) - Esteric(guest) where the steric energies may be evaluated by the molecular mechanics approach.(For the time being I am neglecting the entropic and solvation effects which are not taken into account by a molecular mechanics simulation in the gas phase). The interaction energies are useful in comparing the relative stabilities of different conformations of the complex, for example different orientations of the guest with respect to the host. Now comes my question: given two different complexes of the same host with two different guests (or two hosts with the same guest), is it correct to compare the interaction energies of the two complexes in order to find out which is the better guest (or the better host)? I am not sure that this is correct since, when the two guests (or the two hosts) are not stereoisomers or not even isomers, the two complexes have different chemical compositions. Thanks to all who reply. I will summarize. Paolo =============================================================================== Date: Fri, 23 Feb 96 10:18:25 +0100 From: lnl@novo.dk (Leif Norskov) To: mencarel@axcasp.caspur.it Subject: Re: CCL:Interaction energy in host-guest complex You asked: > is it correct to compare the interaction energies > of the two complexes in order to find out which > is the better guest ? Only if you correct for the difference in "Esteric(guest)". Cubane (CH)8 has a higher internal strain energy than methane, and hence will most likely have a higher Esteric(complex), yet it may have a better fit (interaction) with a host. Regards, Leif Norskov Novo Nordisk A/S Copenhagen Denmark lnl@novo.dk ---------------------------------------------------------------------------- Date: Fri, 23 Feb 1996 10:43:12 +0100 From: Per-Ola Norrby To: "Dr. Paolo Mencarelli" Cc: chemistry@www.ccl.net Subject: Re: CCL:Interaction energy in host-guest complex Dear Paolo, The comparison you want to make is correct, as long as no new covalent bonds are formed between the guest and host. The bond-specific enthalpy contributions will all cancel out. As you stated, you will only get the steric component in the gas phase from this comparison (closely related, but not identical to the enthalpic component). My gut feeling in this is that the most serious error in this would be the neglect of solvation. However, you should also be aware of the risk of entropy-enthalpy compensation: To make it short, the entropy difference will probably give a significant contribution OPPOSITE to that of the enthalpy. For a short summary on this subject, see: J.D.Dunitz, Chemistry & Biology, 2(11), 709-712, 1995. If you are comparing to experimental enthalpies, you mainly have to worry about the solvation part (there are several good computational approaches), but if you are trying to correlate binding constants at one temperature, you will definitely have to take a look at the above article. The really bad part of this is that it's pretty hard to estimate the entropy contribution from a weak interaction computationally. PeO Norrby ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ * Per-Ola Norrby * The Royal Danish School of Pharmacy, Dept. of Med. Chem. * Universitetsparken 2, DK 2100 Copenhagen, Denmark * tel. +45-35376777-506, +45-35370850 fax +45-35372209 * Internet: peon@medchem.dfh.dk, http://compchem.dfh.dk:/ ------------------------------------------------------------------------------- Date: Fri, 23 Feb 96 2:12:19 PST From: Thomas Fox To: "Dr. Paolo Mencarelli" Subject: Re: CCL:Interaction energy in host-guest complex > > Now comes my question: given two different complexes of the same host with > two different guests (or two hosts with the same guest), is it correct to > compare the interaction energies of the two complexes in order to find out > which is the better guest (or the better host)? > I am not sure that this is correct since, when the two guests (or the two > hosts) are not stereoisomers or not even isomers, the two complexes have > different chemical compositions. > Thanks to all who reply. I will summarize. > Ciao Paolo, I guess what you really want to know if you want to investigate the relative binding of different guests to the same host (or vice versa) is the relative free energy of binding \Delta\Delta G. If you draw a free energy cycle: \Delta G1 Guest1 + Host ------------> Guest1 in Host | | | | | | | \Delta G3 | \Delta G4 | | | | | | | | v \Delta G2 v Guest2 + Host ------------> Guest2 in Host the total free energy when you go around in this cycle \Delta G1 + \Delta G4 -\Delta G2 - \Delta G3 =0 Or, rearranged: \Delta G1 - \Delta G2 = \Delta G3 -\Delta G4 = \Delta\Delta G \Delta G1 and \Delta G2 are the experimentally available results (very often only in form of the relative free energy of binding, i.e. \Delta G1 - \Delta G2), whereas from a calculation of whatever sort you get sth. that involves both \Delta G3 and \Delta G4. Some remarks. - just looking at MM energy might get you some idea, but you really want to look at *free energies* instead. Maybe, very maybe if you two extremely similar hosts and/or guests you might get away by looking only at energy contributions and neglecting entropy. Just imagine a guest that is a little bigger than the other which might prohibit some crucial movement inside the host... Only if you - as you obviously have done - have the same guest in different conformations/orientations the approach of looking only at MM energies has some chance to yield meaningful numbers. - be sure to include solvent contribution in some way. I have had situations where my free energy calculations indicated a \Delta G4 that was more or less =0. So the experimentally observed binding energy difference was entirely due to different solvation of my two guests (40 atom haptens to a catalytic antibody that only differed in a CH2 vs a S in a side chain). - my very personal opinion (no flames please): if you cant or dont want to address these two points (*free* energies and solvation), you might as well use your intuition and not bother doing calculations at all. Yes, I know too well that free energy calculaions are expensive and not always feasible due to time constraints...but as usual you only get what you pay for. As I said already above, if you use the same guest in different conformations you might get away with only using the steric energy that you get out of some MM program - after all, the solvation energy is the same, and the entropic contributions are probably almost the same, but when you look at chemically different guests or hosts this approximation is not longer valid. You may want to look at some of the papers from the Gunsteren, Karplus etc group, or also from our group (Kollman) that deal with free energy calculations for protein-ligand binding and host-guest systems. There is a nice book, if I remember right its called Biomolecular Simulations in Liquids or sth. like that, edited by van Gunsteren, that sums up a lot of these questions quite nicely. I think there is also a review by Peter Kollman in the 1993 or 1994 Chemical Reviews that addresses these problems. Ciao and feel free to contact me if you have further questions, Th. -- Thomas Fox Dept. of Pharmaceutical Chemistry University of California e-mail: fox@cgl.ucsf.edu San Francisco, CA 94143-0446 http://www.amber.ucsf.edu/~fox/index.html ,,, (o o) -----------------------------ooo-(^)-ooo------------------------------------- U Wer Kleinstweich Produkte kauft oder in Umlauf bringt wird mit Abhaengigkeit nicht unter zehn Jahren bestraft. ----------------------------------------------------------------------------- #include '../.disclaimer' -------------------------------------------------------------------------------- Date: Fri, 23 Feb 1996 11:46:07 +0000 (GMT) From: Paddy Kane <94970459@vax1.dcu.ie> To: p.mencarelli@caspur.it Subject: Re: CCL:Interaction energy in host-guest complex Hi Paolo, I model the interaction of calixarenes with various types of hosts - neutral and ionic - with molecular mechanics. The nature of molecular mechanics is such that the difference in energies between two conformational isomers is relevant but that the greater the chemical difference in the two molecules being compared the less relevant is the energy difference. Regards, Paddy. Paddy Kane School of Chemical Sciences Dublin City University Dublin 9 Ireland 94970459@vax1.dcu.ie ------------------------------------------------------------------------------ Date: Fri, 23 Feb 96 15:18:10 +0300 From: Igor V. Pletnev To: p.mencarelli@caspur.it Subject: Re: CCL:Interaction energy in host-guest complex Hi Paolo it's undoubtedly correct if you (i) compare deltas not Es (and you do so), and (ii) explicitly say that you are evaluating and comparing not interaction energies but "strain which the complexation costs" or "steric contribution into energy". For many reasons, mol.mech's kilocalories or kiloJoules are not those as in experiment (or even in other computational scheme); however, the former may be in nice, sometimes even linear, correlation with the last. Naturally, you have good chances for such a correlation only if the same binding scheme is operative. Many examples are in the field of host-guest with inorganic guests (e.g., crowns-potassium: Hay et al., JACS, 1993, 11158.) Regards, Igor ------------------------------------------------------------------------------- Date: Fri, 23 Feb 1996 08:56:49 -0500 From: Trevor Creamer To: "Dr. Paolo Mencarelli" Subject: Re: CCL:Interaction energy in host-guest complex On Feb 23, 3:13am, Dr. Paolo Mencarelli wrote: > Subject: CCL:Interaction energy in host-guest complex > Hi all, Hi Paolo, > Now comes my question: given two different complexes of the same host with > two different guests (or two hosts with the same guest), is it correct to > compare the interaction energies of the two complexes in order to find out > which is the better guest (or the better host)? I think perhaps you could come to very erroneous conclusions doing this. Neglecting entropy and solvation effects can lead to serious errors unless the two different guests are very, very similar in size, shape, flexibility and chemical composition (I'll focus on guests, but the same applies to hosts). Let me illustrate with two simple examples: (i) Entropy - Imagine you have a hypothetical host that binds n-pentane. The energy of binding is dE(n-pentane). When bound the pentane is held rigidly in a fairly extended conformation. It has lost all of its configurational entropy. The configurational entropy for rotation about a single C-C bond is roughly 0.7 kcal/mol. Now let's imagine the same host binding 2-pentene. The energy of binding, dE(2-pentene) is going to be very close to that of dE(n-pentane) since they are chemically very similar (at least to a molecular mechanics calculation they are). The solvation component will also be very similar. However, rotation about the double bond is obviously much harder than rotation about a single bond. So, roughly, we can say that rotation about C=C has a configurational entropy of about 0.0 kcal/mol. So the 2-pentene will lose 0.7 kcal/mol less entropy upon binding, which is enough to make it a better guest than n-pentane, but you won't see that if you ignore entropy in your calculations. (ii) Solvation - This one's a little harder to find a clear example for. Ignore entropic effects for the moment. This time imagine a hypothetical host that binds the hydroxyl group of an alcohol. The binding site is deep inside the host. The host will bind 1-pentanol and cyclo-pentanol. When bound, both of these guests are completely shielded from solvent (that way we only have to think about solvent effects on the unbound guests). In water, the 1-pentanol is flexible and, on average, will adopt a relatively extended conformation. Thus, much of the molecule is in contact with solvent. The cyclo-pentanol is much more restrained and, since it is a ring, buries much of its surface against itself (i.e. the inside of the ring is not accessible to solvent). So you can expect the cyclic molecule to have less aliphatic surface area exposed to the water. This would argue that binding the 1-pentanol (removing it from water) should be more favorable (if you ignore entropy). But you won't necessarily see that from a molecular mechanics calculation. I hope this makes sense (I've just got into work and haven't had a coffee yet). Good luck ! Cheers, Trevor -- ________________________________________________________________________ Trevor P. Creamer Dept. of Biophysics Email - trevor@grserv.med.jhu.edu & Biophysical Chemistry Phone - (410) 614 3972 Johns Hopkins University Fax - (410) 614 3971 School of Medicine Baltimore, MD 21205 U.S.A. ________________________________________________________________________ -------------------------------------------------------------------------------- Date: 23 Feb 1996 09:25:58 -0800 From: Tom Hendrickson To: "Dr. Paolo Mencarelli" Subject: Re: CCL-Interaction energy i Reply to: RE>CCL:Interaction energy in host-guest complex I will be discussing this topic at the upcoming meeting in Erice, Scicily in May, "Experimental and Computational Approaches to Structure Based Drug Design" In general, the answer is no because binding may be driven by any number of thermodynamic factors and unless you have a "comprehensive" function you cannot be correct with every case. Also, you are only looking at half the binding equation in detail. This said, however, we have done both molecular dynamics and molecular mechanics minimization and have looked at interaction energies beween protein and ligand. As always, with ligands whose binding is driven primarily by the same physics , correlations can be found between ,for example, between the binding free energies of a series of ligands to FKBP and the average proterin/ligand interaction energies as derived from MD simulations. However, these simulations were not done in the gas phase, where electrostatic effects in solution are poorly modelled. Regards, Tom Hendrickson Agouron Pharmaceuticals,Inc. San Diego, Ca ------------------------------------------------------------------------------ Date: Fri, 23 Feb 96 12:40:02 EST From: Igor Shamovsky To: p.mencarelli@caspur.it Subject: Host-guest complexes. Hi Paolo: Yes, you can definitely do that, of course, if you take into account geometric relaxation of both host and guest, and calculate their specific mutual orientations. Dr. Igor Shamovsky Chemistry Department Queen's University Kingston, Ontario Canada K7L 3N6 ------------------------------------------------------------------------------- Date: Fri, 23 Feb 1996 10:27:51 -0800 From: Ben Burke To: "Dr. Paolo Mencarelli" Cc: burke@cardinal.agouron.com Subject: + 3336 CCL:Interaction energy in host-guest comple [40] Per-Ola Norrby On 23 Feb 1996 02:38:14 Dr. Paolo Mencarelli wrote: > The interaction energy (Delta E) of the host and the guest in a given > complex may be defined as: > Delta E = Esteric(complex) - Esteric(host) - Esteric(guest) > where the steric energies may be evaluated by the molecular mechanics > approach.(For the time being I am neglecting the entropic and solvation > effects which are not taken into account by a molecular mechanics simulation > in the gas phase). Paolo, Do not forget the contributions from (1) electrostatics, and (2) intramolecular "strain" energy. The electrostatics will necessarily have to be balanced by solvation. The "strain" energy will be the intramolecular energy difference between an estimate of the global energy minimum and the energy of the conformer which is bound. This strain terms includes stretch, bend, torsional, electrostatic, solvation and steric terms. For a large guest (such as a protein), I suggest constrained minimization to a local energy minima as the "estimate" of the global energy minimum; for smaller systems, many methods have been described to search conformational space reliably. Best to you, --Ben -- --------------------------------------------------------------------------- Benjamin J. Burke Agouron Pharmaceuticals burke@agouron.com Research Scientist 3565 General Atomics Ct Tel: (619) 622-3018 Computational Chemistry San Diego, CA 92121 FAX: (619) 622-3299 --------------------------------------------------------------------------- ---------------------------------------------------------------------------- ----Date: Mon, 26 Feb 96 14:26 From: Jonathan Baell To: Paolo Subject: host-guest complexes Dear Paolo Your measurement of shape complementarity is a good place to start to estimate likely affinities of various hosts for various guests, since a steric mismatch is energetically very expensive (I think proportional to the sixth power of distance; I'd check this, though) and if "it don't fit, it won't bind" no matter how favourable the other factors. If the synthetic targets are readily accessible/available (unlikely perhaps) your binding data of sterically-matched compounds might be very useful in seeing not only how well theoretical calculations can account for your observations, but in developing better theory. However, I would be concerned if you use your method in a predictive sense for synthetically complicated molecules as one might well find that, if you're looking at binding in aqueous solution, the binding is entropy driven. You may indirectly model entropy effects by inclusion of hydrophobic parameters in your molecular mechanics calculations. When combined with steric/Van der Waals (did the net ever decide whether it was "van" or "Van"?; I cannot remember the conclusion) and electrostatic/H-bonding interactions calculations (with charges preferably calculated semiempirically in the complex also, in case of high polarizability, and at least using a solvent continuum model), you might be on safer ground. The program HINT is a good one to look at with respect to its consideration of hydrophobicity (there is a URL for this), and I think attempts are being made to correlate a HINT score with binding constants, and to introduce the capability of calculating and incorporating potential energies in the case of conformationally flexible ligands. In any case, it seems to me that you need not worry about the point of structural dissimilarity, because you are most interested in the energy penalty involved in your particular compound assuming the binding conformation, compared with its calculated global minimum in the absence of host. This value of energy difference is valid in comparisons between structurally dissimilar compounds, but your force field accuracy in calculating this value may be the limiting factor. Thus I would have thought that while you can't directly compare your potential energies of your complexes with each other in the case of structural difference, you can once you have substracted the potential energy of the relevant ligand in its intrinsically lowest energy conformation. Sincerely Jonathan Baell =============================================================================== Regards, Paolo ********************************************************* * Dr. Paolo Mencarelli * * Dipartimento di Chimica - Universita' La Sapienza * * P.le Aldo Moro, 2 00185 Roma - ITALY * * tel 06-49913697 fax 06-490631 * * E-Mail p.mencarelli@caspur.it * ********************************************************* From owner-chemistry@ccl.net Tue Mar 5 05:32:27 1996 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.7.1/950822.1) id FAA24449; Tue, 5 Mar 1996 05:08:25 -0500 (EST) Received: from aeat.co.uk for jim.henshaw@aeat.co.uk by bedrock.ccl.net (8.7.1/950822.1) id FAA01365; Tue, 5 Mar 1996 05:08:22 -0500 (EST) Received: from ccgate.aeat.co.uk by aeat.co.uk (8.7.1/AEAT-GW-1.3) id KAA04262; Tue, 5 Mar 1996 10:08:00 GMT Received: from ccMail by ccgate.aeat.co.uk (IMA Internet Exchange 1.04b) id 13c12750; Tue, 5 Mar 96 10:07:49 +0000 Mime-Version: 1.0 Date: Tue, 5 Mar 1996 10:06:14 +0000 Message-ID: <13c12750@ccgate.aeat.co.uk> From: jim.henshaw@aeat.co.uk (Jim Henshaw) Subject: Re: CCL:FACSIMILE prog. To: chemistry@ccl.net, Brion Jean Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Description: cc:Mail note part Dear Jean Brion, To find out about FACSIMILE contact Mike Amey AEA Technology Building A50 Winfrith Dorchester Dorset DT2 8DH United Kingdom Tel: +44 (0) 1305 202796 Facs: +44 (0) 1305 202663 e-mail : mike.amey@aeat.co.uk Regards Jim Henshaw ______________________________ Reply Separator _________________________________ Subject: CCL:FACSIMILE prog. Author: Brion Jean at Internet-mail Date: 29/02/96 06:50 Dear netters, Does somebody know where I can find information (address, price, etc...) about FACSIMILE program which is used to simulate kinetic experiments ? Thank you in advance, F. Bohr ************************************************** * Dr F. Bohr * * Laboratoire de Chimie Physique * * Universite de Reims, Faculte des Sciences * * Moulin de la Housse * * B.P. 1039 * * 51687 Reims Cedex 2 * * FRANCE * * Tel: (33) 26-05-32-33 * * Fax: (33) 26-05-31-47 * * E-mail: jean.brion@univ-reims.fr * ************************************************** -------This is added Automatically by the Software-------- -- Original Sender Envelope Address: owner-chemistry@ccl.net -- Original Sender From: Address: jean.brion@univ-reims.fr CHEMISTRY@www.ccl.net: Everybody | CHEMISTRY-REQUEST@www.ccl.net: Coordinator MAILSERV@www.ccl.net: HELP CHEMISTRY or HELP SEARCH | Gopher: www.ccl.net 73 Anon. ftp: www.ccl.net | CHEMISTRY-SEARCH@www.ccl.net -- archive search Web: http://www.ccl.net/chemistry.html From antunez@chem.cinvestav.mx Tue Mar 5 08:32:29 1996 Received: from latina.chem.cinvestav.mx for antunez@chem.cinvestav.mx by www.ccl.net (8.7.1/950822.1) id HAA26385; Tue, 5 Mar 1996 07:47:24 -0500 (EST) Received: by latina.chem.cinvestav.mx (5.x/SMI-SVR4) id AA21273; Tue, 5 Mar 1996 06:47:10 -0600 Date: Tue, 5 Mar 1996 06:47:10 -0600 (CST) From: Sandra Antunez To: Reinaldo Pis Diez Cc: chemistry@www.ccl.net Subject: Re: CCL:Performance of different computers In-Reply-To: <199603041924.QAA04705@biol.unlp.edu.ar> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Transfer-Encoding: QUOTED-PRINTABLE Hi Reinaldo: Try ftp://ftp.cdf.toronto.edu/pub/spectable Regards, Sandra Antunez Departamento de Qu=EDmica CINVESTAV-IPN On Mon, 4 Mar 1996, Reinaldo Pis Diez wrote: > Dear netters >=20 > I need to know the performance of a Pentium, 32 Mb, 150 M= hz > against a PowerStation R6000 43P, 32 Mb too. If someone out there could > help me I'd appreciate very much. >=20 > Thanx in advance, >=20 > = =20 > Reinaldo Pis Diez >=20 >=20 > -------This is added Automatically by the Software-------- > -- Original Sender Envelope Address: pis_diez@nahuel.biol.unlp.edu.ar > -- Original Sender From: Address: pis_diez@nahuel.biol.unlp.edu.ar > CHEMISTRY@www.ccl.net: Everybody | CHEMISTRY-REQUEST@www.ccl.net: Coordin= ator > MAILSERV@www.ccl.net: HELP CHEMISTRY or HELP SEARCH | Gopher: www.ccl.net= 73 > Anon. ftp: www.ccl.net | CHEMISTRY-SEARCH@www.ccl.net -- archive search > Web: http://www.ccl.net/chemistry.html=20 >=20 >=20 From fgonzale@lauca.usach.cl Tue Mar 5 09:35:51 1996 Received: from ralun.usach.cl for fgonzale@lauca.usach.cl by www.ccl.net (8.7.1/950822.1) id JAA27946; Tue, 5 Mar 1996 09:26:19 -0500 (EST) Received: from lauca.usach.cl (lauca.usach.cl [158.170.64.28]) by ralun.usach.cl (8.6.11/8.6.9) with ESMTP id KAA11144 for ; Thu, 5 Mar 1992 10:28:00 -0600 Received: (from fgonzale@localhost) by lauca.usach.cl (8.6.12/8.6.12) id LAA28345; Tue, 5 Mar 1996 11:28:28 -0300 From: Fdo Danilo Gonzalez Nilo Message-Id: <199603051428.LAA28345@lauca.usach.cl> Subject: DESIGN of ANTIBODIES!!!!? To: CHEMISTRY@www.ccl.net Date: Tue, 5 Mar 96 11:28:27 ADT Cc: fgonzale@lauca.usach.cl Dears colleges: Has anybody some references about Design of Catalytic Monoclonal Antibodies for Catalysis of Non-biological Reactions ..?? I will summarize the replies Sincerely Yours, Fernando Danilo Gonzalez N. University of Santiago de Chile casilla 307, Santiago-2, Chile fono: 681 2575 E-mail : fgonzale@lauca.usach.cl fax : (562) 681 2108 *************************************************************************x From rosa@risc2.iaif.pa.cnr.it Tue Mar 5 10:32:30 1996 Received: from risc2.iaif.pa.cnr.it for rosa@risc2.iaif.pa.cnr.it by www.ccl.net (8.7.1/950822.1) id KAA29004; Tue, 5 Mar 1996 10:25:24 -0500 (EST) From: Received: from rosa.iaif.pa.cnr.it by risc2.iaif.pa.cnr.it (AIX 3.2/UCB 5.64/4.03) id AA19415; Tue, 5 Mar 1996 16:24:50 +0100 Date: Tue, 5 Mar 1996 16:24:50 +0100 Message-Id: <9603051524.AA19415@risc2.iaif.pa.cnr.it> X-Sender: rosa@risc2.iaif.pa.cnr.it Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: chemistry@www.ccl.net Subject: AMBER potentials X-Mailer: Dear Netters, I'd like to know if polysaccharides - polysaccharides polysaccharides - polypeptides potentials are available for AMBER. Thanks Dr. Rosetta Noto CNR-IAIF Via Archirafi 36 I-90123 Palermo, ITALY From juanca@daphne.qf.ub.es Tue Mar 5 12:32:31 1996 Received: from daphne.qf.ub.es for juanca@daphne.qf.ub.es by www.ccl.net (8.7.1/950822.1) id LAA01211; Tue, 5 Mar 1996 11:35:51 -0500 (EST) Received: from malva.qf.ub.es by daphne.qf.ub.es with SMTP (1.37.109.4/16.2) id AA23087; Tue, 5 Mar 96 17:33:20 +0100 X-Mailer: InterCon TCP/Connect II 2.2.1 Mime-Version: 1.0 Message-Id: <9603051736.AA14263@malva.qf.ub.es> Date: Tue, 5 Mar 1996 17:36:14 +0100 From: "Juan Carlos Paniagua" To: "Juan Carlos Paniagua" , chemistry@www.ccl.net Subject: Summary: WYSIWYG LaTeX for Mac? Content-Type: Multipart/Mixed;boundary=part_AD622C0D00166B2F00000001 --part_AD622C0D00166B2F00000001 Content-Type: Text/Plain; charset=US-ASCII Content-Disposition: Inline Dear CCL readers, Here is a summary of my recent posting regarding LaTeX for Mac. Thanks for those responding Original posting: > Does anybody know any WYSIWYG text editor for the Mac that > produces LaTeX code (something like wordperfect for the mac, that > allows viewing the codes in a separate window while you are writing)? > Expressionist generates LaTeX code, but it is not practical as > a text editor. Replies: ------------------------------------------------------------------------------ I am not sure how WYSIWYG it is for LaTeX since I use it as an HTML tag editor, but Alpha has been very useful to me. It is available by anonymous ftp from cs.rice.edu/public/Alpha or from Kagi software. From: Abby Parrill ------------------------------------------------------------------------------ You might try "Alpha" editor. It functions like "emacs". Try From: David Weng ------------------------------------------------------------------------------ I don't think such a thing exists. The closest thing that I know of is TeXtures, by Blue Sky Research . It recompiles the TeX document as you type, which is in many ways faster and more convenient than the usual edit-save-compile-view cycle that TeX requires. Still, you have to enter the TeX codes yourself; I guess it's the equivalent of typing into the WordPerfect codes window and having the formatted document updated as you type (but not as fast). The LaTeX codes for entering equations are pretty easy to learn and provide a much faster way of entering equations than Expressionist (which I also used for many years.) That, and the fact that TeX automatically numbers and cross-references your equations, tables, footnotes, and sections for you, are the main things that make it worthwhile in my opinion. Otherwise, it's a pretty poor replacement for a word processor; it's very painful to try to create a document that doesn't closely follow one of the basic styles that are defined for you. From: W. Thomas Pollard ------------------------------------------------------------------------------ There is a commercial product called Textures, but I think it is WYSIWYG TeX instead of LaTeX, but I'm not shure. I myself worked with TeX. From: Blue Sky Research ------------------------------------------------------------------------------ I used to use a package called Textures on the Mac that provided a nice LaTeX implementation. It was a commercial package. I don't recall the distributor, but perhaps you can find it on the Web. From: John Woods or ------------------------------------------------------------------------------ --part_AD622C0D00166B2F00000001-- From pis_diez@nahuel.biol.unlp.edu.ar Tue Mar 5 15:19:15 1996 Received: from biol.unlp.edu.ar for pis_diez@nahuel.biol.unlp.edu.ar by www.ccl.net (8.7.1/950822.1) id OAA06481; Tue, 5 Mar 1996 14:32:21 -0500 (EST) Received: from jubert.biol.unlp.edu.ar (jubert.biol.unlp.edu.ar [163.10.7.25]) by biol.unlp.edu.ar (8.6.9/8.6.9) with SMTP id QAA08211; Tue, 5 Mar 1996 16:20:19 -0300 Date: Tue, 5 Mar 1996 16:20:19 -0300 Message-Id: <199603051920.QAA08211@biol.unlp.edu.ar> X-Sender: pis_diez@nahuel.biol.unlp.edu.ar X-Mailer: Windows Eudora Version 1.4.3 Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="=====================_826067427==_" To: chemistry@www.ccl.net From: pis_diez@nahuel.biol.unlp.edu.ar (Reinaldo Pis Diez) Subject: Summary on Performance of different computers Cc: pis_diez@biol.unlp.edu.ar X-Attachments: C:\WINDOWS\INTERNET\PIS_DIEZ\PERFORMA.NCE; --=====================_826067427==_ Content-Type: text/plain; charset="us-ascii" --=====================_826067427==_ Content-Type: text/plain; charset="us-ascii" Content-Disposition: attachment; filename="PERFORMA.NCE" Dear netters Yesterday I posted the following message concerning the performance of a Pentium vs. an R6000 43P: > > I need to know the performance of a Pentium, 32 Mb, 150 Mhz > against a PowerStation R6000 43P, 32 Mb too. If someone out there could > help me I'd appreciate very much. > And this is the summary (thanks to all who sent me some coordinates to search!!) 1) serge@tiger.jsums.edu (serge krashakov) http://performance.netlib.org/ 2) irisawa@agc.co.jp (Jun Irisawa) http://ccwf.cc.utexas.edu/~zanjal/bench.SPEC.txt 3) Sandra Antunez ftp://ftp.cdf.toronto.edu/pub/spectable 4) "Dr. Heinz Poehlmann" http://www.specbench.org  --=====================_826067427==_-- From pis_diez@nahuel.biol.unlp.edu.ar Tue Mar 5 15:32:33 1996 Received: from biol.unlp.edu.ar for pis_diez@nahuel.biol.unlp.edu.ar by www.ccl.net (8.7.1/950822.1) id OAA06791; Tue, 5 Mar 1996 14:39:41 -0500 (EST) Received: from jubert.biol.unlp.edu.ar (jubert.biol.unlp.edu.ar [163.10.7.25]) by biol.unlp.edu.ar (8.6.9/8.6.9) with SMTP id QAA08217; Tue, 5 Mar 1996 16:20:57 -0300 Date: Tue, 5 Mar 1996 16:20:57 -0300 Message-Id: <199603051920.QAA08217@biol.unlp.edu.ar> X-Sender: pis_diez@nahuel.biol.unlp.edu.ar X-Mailer: Windows Eudora Version 1.4.3 Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="=====================_826067465==_" To: chemistry@www.ccl.net From: pis_diez@nahuel.biol.unlp.edu.ar (Reinaldo Pis Diez) Subject: Summary on Tight Binding program Cc: pis_diez@biol.unlp.edu.ar X-Attachments: C:\WINDOWS\INTERNET\PIS_DIEZ\TIGHT.BIN; --=====================_826067465==_ Content-Type: text/plain; charset="us-ascii" --=====================_826067465==_ Content-Type: text/plain; charset="us-ascii" Content-Disposition: attachment; filename="TIGHT.BIN" Dear netters Yersterday I posted the following mail: > Does anyone know where I can find a free (if possible) code >for doing tight binding calculations? Any reference would be welcome. Basically, there are two available codes for TB calculations: YAeHMOP and EHMACC. The summary follows below. I'd wish to acknowledge these people for the answers. *********************************************************************** 1) tang@Xtended.chem.cornell.edu The extended Huckel type (EHT) of tight-binding program set called YAeHMOP (Yet Another exnteded Huckel Molecular Program) developed by Mr. Greg Lundrum in the Roald Hoffmann group is available for free, and you can down-load it thru WWW or anon FTP. This program set consists of BIND program for tight-binding as well as molecular EHT calculations and of VIEWKEL for graphic display of orbital interaction diagram, band structure, DOS, COOP, and much more. It is very user-friendly, and it works on Unix X-Window systems as well as on Mac (I believe it's also available now). For more info, please take a look at http://overlap.chem.cornell.edu:8080/yaehmop.html. On this web site, you can down load the program manual. *********************************************************************** 2) Francois Savary I guess there are only few free code, the ones I know are the EHMACC Program (from our group) which is not yet available and the one of Hoffmann's group YaEHMOp which you can find at the following URL: http://overlap.chem.cornell.edu:8080/~landrum/yaehmop.html *********************************************************************** 3) antonio@linus.qui.ub.es (Antonio Buljan) ... Antonio also cites the two packages mentioned above.  --=====================_826067465==_-- From schrecke@zinc.chem.ucalgary.ca Tue Mar 5 15:34:37 1996 Received: from zinc.chem.ucalgary.ca for schrecke@zinc.chem.ucalgary.ca by www.ccl.net (8.7.1/950822.1) id OAA06956; Tue, 5 Mar 1996 14:58:58 -0500 (EST) From: Received: by zinc.chem.ucalgary.ca (AIX 3.2/UCB 5.64/4.03) id AA16492; Tue, 5 Mar 1996 12:48:31 -0700 Message-Id: <9603051948.AA16492@zinc.chem.ucalgary.ca> Subject: Conservation of Difficulty To: chemistry@www.ccl.net Date: Tue, 5 Mar 1996 12:48:29 -0700 (MST) Reply-To: schrecke@zinc.chem.ucalgary.ca (Georg Schreckenbach) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Hi everybody, a couple of years ago, one of my theoretical physics professors cited the "Law of Conservation of Difficulty". I thought I should share this law with the computational chemistry community on the net. As the name suggests, the law states that the difficulty of a problem is conserved, no matter how you reformulate it. I will demonstrate this by a few examples. Take Density functional theory (DFT). We start off with the terribly complicated Schroedinger equation -- for the electrons in a molecule, say. We reformulate it in a very clever way to obtain DFT. Now what have we got? Indeed, we have a beautiful formulation of the same problem: the basic variable is the density, an observable that depends on three coordinates, rather than 3*N as is the case with the wave function. Thus, the problem has been simplified considerably. However, all the difficulty comes back in the exchange-correlation functional. (Remember that its functional form is unknown). Another example is given by Molecular Mechanics. Again, the very difficult problem of the time-dependent Schroedinger equation is reformulated as simple classical equations of motion. However, the difficulty is conserved. In this case, it pops up in the necessity to obtain reliable force fields. I suppose I have to modify the law somewhat since it is certainly possible to make life MORE complicated (by doing lots of stupid things). Maybe the "difficulty" is an entropy-like property? Does anybody want to comment on the above? If so, then I shall summarize to the net. In particular, I would like to get a reference ... Yours, Georg P.S. Don't take me too serious on this one ... ============================================================================== Georg Schreckenbach Tel: (Canada)-403-220 8204 Department of Chemistry FAX: (Canada)-403-289 9488 University of Calgary Email: schrecke@zinc.chem.ucalgary.ca 2500 University Drive N.W., Calgary, Alberta, Canada, T2N 1N4 ============================================================================== From jerry@dft.chem.cuhk.hk Tue Mar 5 19:32:33 1996 Received: from iris.chem.cuhk.hk for jerry@dft.chem.cuhk.hk by www.ccl.net (8.7.1/950822.1) id TAA08813; Tue, 5 Mar 1996 19:05:17 -0500 (EST) Received: from dft.chem.cuhk.hk by iris.chem.cuhk.hk via SMTP (931110.SGI/940406.SGI.AUTO) for chemistry@www.ccl.net id AA21381; Wed, 6 Mar 96 08:04:30 +0800 Received: by dft (950511.SGI.8.6.12.PATCH526/930416.SGI.AUTO) id IAA07074; Wed, 6 Mar 1996 08:04:24 +0800 Date: Wed, 6 Mar 1996 08:04:23 +0800 (HKT) From: Jerry C C Chan To: chemistry@www.ccl.net Subject: Shielding is a ground state property Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear Netters, After going through some articles, I have a thinking about the future aspect of the calculation of shielding constant and I would like to have your comments. Currently, magnetic effect on a molecule is treated via perturbation theory in density functional approach. The calculation of shielding constant involves excited states as in CHF theory. Very recently it is published the theoretical framework for incorporating either the magnetic field (suggested by C.J. Grayce and R.A. Harris at Berkeley) or the current density (by G. Vignale and M. Rasolt) into the exchange-correlation (XC) functional. Although an explicit XC functional which can be used for the magnetic-field density functional theory or the current density functional theory is not yet realized, the equations of shielding calculation for both theory have been published (Grayce and Harris for MDFT; Colwell and Handy at Cambridge for CDFT). My speculation: Once the XC functional for either theory appears, diamagnetic shielding would become an eigenvalue problem [assume the guage origin is chosen at the nucleus]. The paramagnetic shielding constant is just evaluated in the same way as the diamagnetic term in CHF theory, i.e. evaulation of the triple product of the ground state bra, the operator of the vector potential of the electrons exerted by the magnetic moment of the nucleus and the ground state ket. It means that the time required for shielding calculation in MDFT or CDFT approach is of the same order of magnitude as the diamagnetic terms in CHF, provided that the SCF density can be evaluated as easy as in HF approach. Since excited state is no longer required for the calculation, shielding constant should then become a ground state property. Accuracy of the shielding results depend solely on the quality of the XC functional. Similar sayging is applicable to all second order observables with respect to magnetic field. Cheers, Jerry C.C. Chan From jerry@dft.chem.cuhk.hk Tue Mar 5 19:46:54 1996 Received: from iris.chem.cuhk.hk for jerry@dft.chem.cuhk.hk by www.ccl.net (8.7.1/950822.1) id TAA09024; Tue, 5 Mar 1996 19:30:46 -0500 (EST) Received: from dft.chem.cuhk.hk by iris.chem.cuhk.hk via SMTP (931110.SGI/940406.SGI.AUTO) for chemistry@www.ccl.net id AA21402; Wed, 6 Mar 96 08:30:10 +0800 Received: by dft (950511.SGI.8.6.12.PATCH526/930416.SGI.AUTO) id IAA07146; Wed, 6 Mar 1996 08:30:08 +0800 Date: Wed, 6 Mar 1996 08:30:08 +0800 (HKT) From: Jerry C C Chan To: chemistry@www.ccl.net Subject: Summary: DFT package for chemical shielding Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear Netters, I would like to express my gratitude to all who response to my question: what commercial available softwares which can be used to calculate the chemical shielding constants of transition metal using Density Functional Theory method, in addition to Gaussian94 and deMon-NMR1p0? Below please find some of the suggestions: The UniChem package available from Cray can compute NMR chemical shifts and shielding tensors using DFT. The method is similar to that published by Salahub for his 'uncoupled' approximation and uses either IGLO or LORG. Results from this method are generally pretty good for C; experience with other elements is limited. ======================================================= George Fitzgerald Cray Research, Inc. george@cray.com Chem Manager 655E Lone Oak Dr. 1-612-683-3676 Eagan, MN 55121 1-612-683-3099 (fax) ======================================================= There are a few codes that can routinely do nmr shielding calculations using DFT. In addition to Gaussian, you might want to look at Turbomole, available from Biosym Software (now MSI), San Diego; Cadpac6, available on request from Roger Amos, Cambridge - email rda@uk.ac.cam.ch.theor; DGuass3.0, part of the Cray Soft Unichem package. ############################################################################## Phillip Sinclair Royal Institution of Great Britain Phone: +44 (0)171-409-2992 21 Albemarle Steet Fax: +44 (0)171-629-3569 LONDON W1X 4BS, UK email: phillip@ri.ac.uk ############################################################################## You mentioned in your posting G94 and deMon. I doubt that there is any other commercially available DFT-NMR code (yet). I am currently working on a similar implementation project that is based on the Amsterdam code ADF. Eventually, it will for sure end up in the commercially available version of ADF, but not in the very near future. In any case, please summarize your findings! ============================================================================== Georg Schreckenbach Tel: (Canada)-403-220 8204 Department of Chemistry FAX: (Canada)-403-289 9488 University of Calgary Email: schrecke@zinc.chem.ucalgary.ca 2500 University Drive N.W., Calgary, Alberta, Canada, T2N 1N4 ============================================================================== From young@argus.cem.msu.edu Tue Mar 5 21:32:35 1996 Received: from argus.cem.msu.edu for young@argus.cem.msu.edu by www.ccl.net (8.7.1/950822.1) id VAA10512; Tue, 5 Mar 1996 21:02:50 -0500 (EST) From: Received: from slater.cem.msu.edu by argus.cem.msu.edu via ESMTP (950215.SGI.8.6.10/940406.SGI) for id VAA18279; Tue, 5 Mar 1996 21:03:36 -0500 Received: by slater.cem.msu.edu (950215.SGI.8.6.10/) for chemistry@www.ccl.net id VAA29033; Tue, 5 Mar 1996 21:04:24 -0500 Date: Tue, 5 Mar 1996 21:04:24 -0500 Message-Id: <199603060204.VAA29033@slater.cem.msu.edu> To: chemistry@www.ccl.net Subject: RE: Conservation of Difficulty Georg writes: > > a couple of years ago, one of my theoretical physics > professors cited the "Law of Conservation of Difficulty". > I thought I should share this law with the computational > chemistry community on the net. > > As the name suggests, the law states that the difficulty > of a problem is conserved, no matter how you reformulate it. > I will demonstrate this by a few examples. I will agree that there are very often trade offs in the difficulty of methods. However, I will not admit to any sort of law of conservation. Let me give a counter example. Consider the solution of the Schrodinger equation for the hydrogen atom. If we did not know the exact solution we could put incredible amounts of work into extremely accurate calculations using DFT, GTO expansions, cubic splines, polynomials, etc. Now consider how difficult hydrogen atom calculations could be even knowing the exact solution. The logarithm and exponential functions have proven so useful that they have been built into calculators and programming languages and reliably return results to the precision of the machine. If this were not the case, we would have to spend quite a bit of time making sure that we are accurately computing exponentials and logarithms, or we could have an enormous database holding a table of logarithms ( 60 years ago every scientist and engineer had a table of logarithms handy at all times ). Now extrapolate from our current state of theory. If we knew the exact analytic solution to the Schrodinger equation for molecules most of what we do now would be trivial on the smallest PC. This would still not prevent computational chemists from eating every bit of computer power in sight as they tried to either deal with relativistic effects or the transport of drugs through cell wall membranes. Although computers are incredibly powerful tools that I would not want to live without, they have also made us lazy. In the past such scientific problems would have either been shelved until the mathematicians had come up with a technique for solving it or scientists would have worked on it until a solution was found. Although numerical techniques allow us to jump past the development of mathematics, sooner or later we will have to go back and get the original problem right. I firmly believe that the Schrodinger equation will one day be as easily dealt with as trignometric functions on a calculator. Now ask your self two questions. 1. How important would it be to work on an analytic solution to the Schrodinger equation? 2. What would be the chances that you could get funding for this project? I will let you draw your own conclusions. Dave Young young@slater.cem.msu.edu -------------------------------------------------------------------------- No assumption or approximation is reasonable for all cases. Corollary: Assumptions must be both rationalized and checked. But it is more important to check them. --------------------------------------------------------------------------