From chemistry-request@ccl.net Fri Mar 6 12:07:24 1992 Date: Fri, 6 Mar 92 08:39:02 -0600 From: berkley@wubs.wustl.EDU (Berkley Shands) Subject: Re: Inquiry for cycle-burners To: chemistry@ccl.net Status: R You can burn many cycles looking for small molecule active analogs, and doing NMR constraint calculations. As in CPU months :-) Berkley From chemistry-request@ccl.net Fri Mar 6 13:51:10 1992 Date: Fri, 6 Mar 92 16:22:25 +0100 From: martin@link.sunet.SE (Martin Norin, Dept. Biochem., Royal Inst. Technol., Subject: MOPAC ESP CHARGES again... To: "chemistry@ccl.net"@kth.sunet.SE Status: R Dear Netters, I sent out an note on strange charges using the MOPAC60_ESP program. Some interesting points was reflected to me. They are summarized below But first I have to mention that I think I've isolated the problem of the completely enormous ESP charges that I've got (-66 , +13 etc). The key of the problem is that I used (for the first time) an input file with multiple input data concatenated after each other. The result is that one might run several jobs after eachother in a pseudo-que. Un- fortunately this gives a problem with the ESP charges. The program stores all ESP points during the whole calculation. The NUMBER OF POINTS reported in the .out file grows from one molecule (conformation) to another in the same concatenated run. Also the "old" points seems to be taken into account when calculating the ESP charges for the new molecule. When doing exactly the same runs but with separate input files the problem with strange charges disappear. Thank you all for comments and suggestions about this. Here follows a summary of replies: ------------------------------------------------------------------------ Dr Carlos H. Faerman writes: One thing you could do is to convert the observation points where you calculate the electrostatic potential at into a 'pdb'like file In other words display all the points you are using to calculate the potential and carefully monitor the distances between these points and the atomic positions ... I would not be surprised if one of these distances was very close to an atom specially when you go from one conformer to another one. This might give you an answer to your problem. --------------------------------------------------------------------- My comment: THANK YOU CARLOS for forcing me to really monitor the points ! -------------------------------------------------------------------- ---------------------------------------------------------------------- Asst. Prof. Leonore Findsen writes: My situation is related to yours. I have been running Monte Carlo conformational searches on molecules with charges of +1. I also have the problem that not all of the charge distributions are the same for the conformations. That problem is mostly solved, but not all of my charges add up to 1! I have been corresponding with one of the authors of the program, Brent Besler (bbesler@vela.acs.oakland.edu) and it seems that the program was parameterized for neutral molecules. At least with my molecules, the charges will not be reasonable unless you perform a MO geometry optimiation of the classical mechanics structures. Slight differences in conformation can lead to large problems in the charge distribution. But then my molecules are _very_ strange. -------------------------------------------------------------------------- Dr Chris Naylor writes: In response to your note on the CompChem Bulletin Board, we use MOPESP fitted charges regularly too and as a general rule have had good experiences with the program. The only relevant piece of information I can relate is that we get completely unbelievable charges as well if we use the MULLIK keyword during the calculation. For reasons that are not known to me, the ESP and MULLIK keywords appear to be incompatible and result in "partial" charges of more than +/- 20 units. I have not seen this documented anywhere, and wonder if it, or something similar, may be responsible for your problem too. -------------------------------------------------------------------------- My comment: In the latest QCPE-bullentin it is mentioned that the MULLIK keyword cause corruption of the .arc output file, while the .out file remains uncorrupted. Maybe this problem is also related to the MULLIK-ESP problem. Also the keyword GRAPH cause this problem. -------------------------------------------------------------------------- ************************************************************************ Martin Norin tel: +46-8-790 7512 Dept. Biochem fax: +46-8-723 1890 KTH e-mail: martin@physchem.kth.se 100 44 Stockholm Sweden ************************************************************************ DISCLAIMER: There is a theory that says that if anyone ever understands the laws of the Universe and why it exists it will immediately be replaced by something even more incomprehensible. Another theory says that this have already happened. /Freely after Douglas Adams. From chemistry-request@ccl.net Fri Mar 6 16:25:19 1992 Date: Fri, 6 Mar 92 13:17:51 EST From: rs0thp@rohmhaas.COM (Dr. Tom Pierce) Subject: Re: Inquiry for cycle-burners To: chemistry@ccl.net Status: R RE: > Joe Leonard > This is sort of a followup to jkl's posting regarding "what's next..." > > I'm interested in learning what folks burn SCADS of computer time doing. > I know they do ab initio QM, as well as large-scale MD simulations, but > is there anything else that's not getting the press? I'd really appreciate > answers from folks managing supercomputer resources, or with large mini- > super's or big superscalar workstation clusters... > I used to think that ab initio calcs used alot of CPU until I discovered conformational analysis of floppy molecules/polymers and equilibrium bulk/solvated systems. These can be studied with monte carlo methods with lots of samples running independently on every computer in sight. Sometime I run Monte Carlo MOPAC calculations and generate boltzmann distributions of interaction energies too. I have found that for my polymer systems Monte Carlo computations are more efficient than huge MD calculations. Most of these calculations are run on clusters of RS/6000s and Alliant minisupers. Monte Carlo approaches are often simple to implement using a network of computers. -- -------------------------------------------------------------------- Sincerely, Thomas Pierce Computer Applications Research | rs0thp@rohmhaas.com Internet Bldg 64C, Rohm and Haas Co. | rs0thp@rohvm1 Bitnet P.O. Box 219 | (215)-785-8989 Voice Bristol, PA 19007 | (215)-781-4204 Fax Official Disclaimer: "The opinions expressed are those of the writer and not the Rohm and Haas Company." From chemistry-request@ccl.net Fri Mar 6 18:12:52 1992 Date: Fri, 6 Mar 1992 14:59 CST From: "JORGE M. SEMINARIO" Subject: nitramide help To: chemistry@ccl.net Status: R Does any of you knows the experimental dissociation energy of nitramide into NO2 and NH2 radicals ? 2HNNO2 ---> NH2 + NO2 Has it been measured ? Your help is appreciated. Regards, Jorge M Seminario Bitnet%"jsmcm@uno" Department of Chemistry Internet%"seminari@cpwsca.psc.edu" University of New Orleans TEL: 504-286-7216 New Orleans, LA 70148 FAX: 504-286-6860 --------------------------------------------------------------- From chemistry-request@ccl.net Fri Mar 6 19:13:53 1992 Date: Fri, 6 Mar 92 14:24:55 EST From: shenkin@avogadro.barnard.columbia.EDU (Peter S. Shenkin) Subject: Re: Inquiry for cycle-burners To: berkley@wubs.wustl.EDU, chemistry@ccl.net (Berkley Shands) Status: R Let's not forget the cycles burned reading this newsgroup! Non-trivial with a fancy gui.... -P. :-) ************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb************************* Peter S. Shenkin, Department of Chemistry, Barnard College, New York, NY 10027 (212)854-1418 shenkin@avogadro.barnard.columbia.edu shenkin@cunixf.BITNET ********** "I've got algorithm -- who could ask for anything more?" ********** From chemistry-request@ccl.net Fri Mar 6 22:03:51 1992 Date: Fri, 6 Mar 1992 18:23 CST From: CMSWALTERS@minna.iit.EDU Subject: looking for info on polyamines To: chemistry@ccl.net Status: R Has anyone done computational studies on polyamines (e.g., putrescine, spermidine, spermine)? Thanks for your help. >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > D. Eric Walters > > Assoc. Prof. Biological Chemistry > > Chicago Medical School phone: 708-578-3000, ext 498 > > 3333 Green Bay Road fax: 708-578-3240 > > North Chicago, IL 60064 email: cmswalters@minna.iit.edu > >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>