From raman@bioc01.uthscsa.edu Wed Jun 9 18:17:38 1993 From: raman@bioc01.uthscsa.edu (C.S.RAMAN) Message-Id: <9306100417.AA08874@bioc01.uthscsa.edu> Subject: Re: DNA software To: rpearls@hawk.dcrt.nih.gov (Robert Pearlstein) Date: Wed, 9 Jun 1993 23:17:38 -0500 (CDT) Robert: Regarding the DNA software, you might want to contact either Dr. Helen Bermen's or Dr. Wilma Olson's group at Rutgers University. They have been working on DNA structure related problems for quite some time and could provide details. Hope this helps -raman -- C.S.Raman raman@bioc01.uthscsa.edu - Internet UNIX Programming & Administration 70412.2354@compuserve.com - CIS SPARC & SGI Systems raman@hermes.chpc.utexas.edu - CHPC Department of Biochemistry craman@launchpad.unc.edu UTHSCSA 7703 Floyd Curl Dr. (210) 567-6623 [Tel] San Antonio, TX 78284-7760 (210) 567-6595 [Fax] ****************************************************************************** If a man's wit be wandering, let him study the Mathematics -Francis Bacon ****************************************************************************** From chp1aa@surrey.ac.uk Thu Jun 10 06:50:18 1993 From: Mr Andrew D Allen Message-Id: <9306101033.AA19743@central.surrey.ac.uk> Subject: SUMMARY: Solvation Methods and Modelling. To: chemistry@ccl.net Date: Thu, 10 Jun 93 11:33:31 BST Its a little overdue, but here is the promised summary of solvation... Original request: >Dear All, > I am interested in the modelling of solvated systems, especially those >of a biological nature. I have had several pointers already as to how I should >go about getting realistic data. > First is to use a molecular mechanics approach, by putting my molecule >in a box and adding solvent to it, adding solvent to my molecule and doing some >kind of MD run in a periodic box. This seems ok but how would MM methods handle >such effects as hydrophobicity, cavity's and solvent accessible surface area. > The second approach I have considered is to use a package called AMSOL >this used a parameterised AM1 method and seems quite extensive, my question to >the net is has anyone done any extensive work with AMSOL or any other solvation >methods. If so I would be glad to hear from you, I will summerise the pro's >and con's of the various methods for the net. > >ANDY ################################################################################ Hi!, About your question on Amsol, check the following reference: "Comparison Methods to Estimate the Free Energy of Solvation: Importance in the modulation of the affinity of 3-Benzazepines for the D1 Receptor". I.Alkorta, H.O.Villar, and J.J.Perez Jour.Comp.Chem Vol.14(5), 620-626 (1993). This paper include a detailed comparison between Delphi and Amsol calculations over a long list of neutral and ionic species. Hope this helps. R.E.Cachau ################################################################################ Andrew, > The second approach I have considered is to use a package called AMSOL this used a parameterised AM1 method and seems quite extensive, my question to the net is has anyone done any extensive work with AMSOL or any other solvation methods. If so I would be glad to hear from you, I will summerise the pro's and con's of the various methods for the net. > Great program, AMSOL . . . Best wishes -- Chris P.S. PM3 is also available with the SM3 solvation model in the currently available (through QCPE) AMSOL 3.0.2. P.P.S. For a less effusively biased opinion than my own, you might take a look at a recent article in J. Comp. Chem. (14 [1993] 620) by Alkorta, Villar and Perez. Our own work has produced a number of articles; I would be happy to send you reprints should you be so inclined. Christopher J. Cramer University of Minnesota Department of Chemistry 207 Pleasant St. SE Minneapolis, MN 55455-0431 (612) 624-0859 ################################################################################ >> Some more references kindly supplied by Chris Cramer. Cramer, C. J.; Truhlar, D. G. "Molecular Orbital Theory Calculations of Aqueous Solvation Effects on Chemical Equilibria" J. Am. Chem. Soc. 1991, 113, 8552. Cramer, C. J.; Truhlar, D. G. "PM3-SM3: A General Parameterization for Including Aqueous Solvation Effects in the PM3 Molecular Orbital Model" J. Comput. Chem. 1992, 13, 1089. Cramer, C. J.; Truhlar, D. G. "Comparative Analysis of the AM1-SM2 and PM3-SM3 Parametrized SCF Solvation Models for Free Energies in Aqueous Solution" J. Comput.-Aid. Mol. Des. 1992, 6, 629. Cramer, C. J.; Truhlar, D. G. "Polarization of the Nucleic Acid Bases in Aqueous Solution" Chem. Phys. Lett. 1992, 198, 74. Cramer, C. J.; Truhlar, D. G. "What Causes Aqueous Acceleration of the Claisen Rearrangement?" J. Am. Chem. Soc. 1992, 114, 8794. Cramer, C. J. "Anomeric and Reverse Anomeric Effects in the Gas Phase and Aqueous Solution" J. Org. Chem. 1992, 57, 7034. Cramer, C. J.; Truhlar, D. G. "Quantum Chemical Conformational Analysis of Glucose in Aqueous Solution" J. Am. Chem. Soc. in press. (Scheduled for June 30 issue). We offer the level of support you might expect for free . . . friendly, but not necessarily reliable in an immediate sense. However, the AMPAC and AMSOL manuals are reasonably lucid. Good luck! Chris >> The manual is lucid, and the references are very helpful. ################################################################################ Maybe you should thank of using molecular dynamics, possible doing free energy calculations. I have no experience with AMSOL, but if you did free energy calculations you could calculate hydration free energies. You can also use normal MD and look at static and dynamic solvation effects on the solvent. Scott ################################################################################ Andrew Allen. Hi, Here at Columbia we have done a lot of work with continuum solvation models in a molecular mechanics framework. We do very well with predictions of free energies of solvations for a range of small organic and bio-organic molecules. We also have success with prediction of free energies of host-guest binding in solvent and things like predicting hydrogen bonding populations for small amides in organic solvents. The main reference for the solvation model is J. Am. Chem. Soc. 112, 6127, 1990. The model is called GB/SA [Generalized Born/ Surface Area] and has both terms which take into account the cavity and Van der Waals terms (proportional to surface area) and polarization terms. And now the advertisment: MacroModel is a modelling package which runs on SGI and IBM RS/6000 workstations. We support versions of AMBER, OPLS/A , MM2 and MM3 forcefields with a number of locally added extensions. A range of molecular dynamics and minimization procedures are available. The program has a graphical front end which allows generation and analysis of molecular structures. Cost is $US750 for academics. Let me know if you want more information. Quentin McDonald. ################################################################################ Hi andy, I'm rsponding to your list posting re: solvated systems. Two things: (1) AMSOL calculations will almost certainly be FAR less computationally intensive than MD simulations, at least for solvation free energies. I have not done this personally, but I have inside sources and have discussed this point with Don Truhlar in the past. I don't know if AMSOL has been used on really large systems yet though. (2) I'm no expert, but I believe that if one had a perfect force field, an infinite number of waters, and infinite computer time, hydrophobicity, solvent-accessible surface area effectsm and cavities would all automatically be taken into account in MD. The task would be to interpret the simulation in terms of these effects. I could be wrong on this point, but this is my present understanding. Of course, the force field alone can be a problem...but there's a lot of precedent in the lit for these simulations. I hope this helps you, best, robert ################################################################################ Andy, I have been using Amsol 3.0.1 for quite a while. The following experience arises using DERINU (numerical derivatives turned on, i.e. full geometry optimization in the AM1 SM2 method) is very time consuming - I left jobs to run for weeks (!) on the SGI 380, and they would not finish; moreover, there is an apparently system dependent bug, in the sense that the DUMP routine does not work, i.e. it does NOT write a restart file every hour (actually it's rewriting function never worked...). The approach I used is optimize in cavity with molecular mechanics, than use AM1 (gas) to get hof, then use AM1 SM2 to get hof + ^G, the substract and get a reasonable ^G for that particular conformer. I ran this for a series of 78+ molecules. Same geometries using Delphi, the solvation model from Insight/Discover (Biosym). While Delphi took less than a day for the whole series, Amsol took weeks. The results were correlated with r > .8 (I'll let you judge what that means and what you want, speed or accuracy). On the other hand, Ampac 4, distributed by Dr. Andy Holder from Kansas City, has an improved version of Amsol that runs 50 to 100 times faster (but does not come with QCPE, you have to pay for it). I am certain you'll find this piece of the puzzle useful. Cheers, Tudor Oprea Research Associate @ Wash U Ctr for Mol Design tudor@wucmd.wustl.edu St.Louis MO >> Agreement on the fact that AMSOL takes quite a while to do it's solvation >> runs. But then if you really want the results to be good then you need to >> have a little patience. I've been trying to get meaningful data for several >> months using MD, amsol has already given the geometry expected for the system >> i'm interested in, and that's just gas phase. Hopefully the Solvation run >> is going to be even closer to the expected conformation. ################################################################################ Thanx for your patience, and if anyone has any further questions or input then I would be glad to here from you. ANDY. From pattabiraman@lsm.nrl.navy.mil Thu Jun 10 05:44:29 1993 Date: Thu, 10 Jun 1993 09:44:29 EDT From: "N. PATTABIRAMAN 202-767-0657" To: chemistry@ccl.net Message-Id: <0096DCFC.4D7B6480.18630@lsm.nrl.navy.mil> Subject: Computational Chemistry I would like to register my name in the computational chemistry group. N. Pattabiraman Code 6030 Naval Research Laboratory Washington DC 20375 Phone Voice: (202)-767-0657 FAX : (202)-767-6874 e-mail : pattabiraman@lsm.nrl.navy.mil Is it possible to get the list of names who have already registered? Please send me e-mail Thanks From UDIM018%FRORS31.BITNET@phem3.acs.ohio-state.edu Thu Jun 10 10:23:39 1993 Date: Thu, 10 Jun 1993 14:23:39 -0400 (EDT) From: UDIM018%FRORS31.BITNET@phem3.acs.ohio-state.edu Subject: Electronic Journals? To: chemistry@ccl.net Message-Id: <01GZ7HMBMB8I8WWNR1@phem3.acs.ohio-state.edu> RS4gTS4gRVZMRVRIDQpEeW5hbWlxdWUgZGVzIEludGVyYWN0aW9ucyBNb2xlY3Vs YWlyZXMNClVuaXZlcnNpdGUgUGllcnJlIGV0IE1hcmllIEN1cmllDQo0IFBsYWNl IEp1c3NpZXUsIFRvdXIgMjIsIFBhcmlzIDc1MDA1DQozMy0xLTQ0LTI3LTQyLTA4 ICh3b3JrKSwgMzMgPSBGcmFuY2U7IDEgPSBQYXJpcw0KMzMtMS00NS00OC02Ny0y MCAoaG9tZSkNCkZBWCAzMy0xLTQ0LTI3LTQxLTE3IChsYWIpOzQ0LTI3LTM4LTY2 KFVuaXZlcnNpdHkpDQplLW1haWwgVURJTTAxOCBhdCBGUk9SUzMxLkJJVE5FVA0K DQpJIGp1c3QgcmV0dXJuZWQgZnJvbSB0aGUgVVNBIGR1cmluZyB0aGF0IHZpc2l0 IG9uZSB5b3VuZyByZXNlYXJjaGVyDQphc2tlZCBtZSBteSBvcGluaW9uIGFib3V0 IGVsZWN0cm9uaWMgam91cm5hbHM/IEkgd291bGQgbGlrZSB0bw0Kc3RpbXVsYXRl IHBlb3BsZSB0byB0aGluayBzZXJpb3VzbHkgYWJvdXQgdGhpcyBwb3NzaWJpbGl0 eS4NClRoZSBvYnZpb3VzIG5lZ2F0aXZlIGZlYXR1cmUgb2Ygc3VjaCBhICJqb3Vy bmFsIiBpcyByZWZlcmVlaW5nDQphbmQgZ29pbmcgdGhyb3VnaCBwZWVyLWdyb3Vw IHJldmlldy4gTW9zdCB5b3VuZyBwZW9wbGUgYXJlIGNvbmNlcm5lZA0KYWJvdXQg YWNjdW1tdWxhdGluZyBhIHB1YmxpY2F0aW9uIHJlY29yZCBmb3IgcHJvbW90aW9u IHJlYXNvbnMgYW5kDQp3b3VsZCBub3QgbGlrZSB0aGVpciB3b3JrIHB1Ymxpc2hl ZCBpbiBhIHJldmlldyBoYXZpbmcgYSBwb29yDQpyZXB1dGF0aW9uLiBQcm9tb3Rp b24gYW5kIG9idGFpbmluZyByZXNlYXJjaCBncmFudHMgYXJlIHNlcmlvdXMNCmJ1 c2luZXNzIGFuZCBzb21ldGltZXMgaW50ZXJmZXIgd2l0aCAicHVyZWx5IiBzY2ll bnRpZmljIGNvbnNpZGVyYXRpb25zLg0KQ3JlYXRpbmcgYSBwcm9mZXNzaW9uYWxs eSBjcmVkaWJsZSBlLWpvdXJuYWwgaGFzIHNvbWUgbGFyZ2UgaHVyZGxlcw0KdG8g b3ZlcmNvbWUuIEJVVC0tDQoNCkluc3BpdGUgb2YgdGhlc2UgYWJvdmUgbmVnYXRp dmlzbXMsIGl0IHNlZW1zIHRvIG1lIHRoZSBlLWpvdXJuYWwNCmlkZWEgaGFzIHNv bWUgZGlzdGluY3QuIFRoZSBBQ1Mgam91cm5hbHMgYXJlIGFscmVhZHkNCmdvaW5n IHRvIHN1cHBsZW1lbnRhcnkgbWF0ZXJpYWwgYXZhaWxhYmlsaXR5IGJ5IGZ0cCwg YXQgYSBwcmljZS4NCkFDUy1vbmxpbmUgaXMgYXZhaWxhYmxlIOJCdCBhbiBoaWdo IHByaWNlKS4gVGhlIGF2YW50YWdlIG9mIGENCnRydWUgZS1qb3VybmFsIHdvdWxk IGJlIHJhcGlkIHJldmlldyBhbmQgcHVibGljYXRpb24uIExvY2FsIHByaW50DQpv dXQgaXMgcG9zc2libGUgaWYgY29ycmVjdGx5IGZvcm1hdGVkLiAgQnV0IG1vcmUg aW1wb3J0YW50IG9mIGFsbA0KaXMgdGhlIHBvc3NpYmlsaXR5IG9mIHVzaW5nIGdy YXBoaWMgY29uZmlndXJhdGVkIGZpbGVzLCB0aGlzDQpwZXJtaXR0aW5nIG9uZSB0 byBsb29rIGF0IGEgcGFydGljdWxhciBzdHJ1Y3R1cmUgZnJvbSB2YXJpb3VzDQpk aXJlY3Rpb25zLiAgSSBwZXJzb25hbGx5IGRvIG5vdCBnZXQgbXVjaCBvdXQgb2Yg bXVjaCBvZiB0aGUgY3VycmVudA0KcHJpbnRlZCBncmFwaGljcyBvZiBhIGNvbXBs aWNhdGVkIHN0cnVjdHVyZS4gU2V2ZXJhbCBqb3VybmFscyByZWNlbnRseQ0Kc3Rh cnRlZCBwcmVzZW50IHRoaXMgc3R1ZmYsIGJ1dCBmaW5kIGl0IHVzZWxlc3MuIElu IGFkZGl0aW9uLA0KaXQgd291bGQgYmUgbmljZSB0byBoYXZlIGNyeXN0YWxsb2dy YXBoaWMgZGF0YSBkaXJlY3RseSBtb3VudGFibGUNCm9uIG9uZSdzIHN0YXRpb24g d2l0aG91dCBoYXZpbmcgdG8gcmV0eXBlIHRoZSBkYXRhIGZyb20gdGFibGVzDQpv ciBnZXQgaWYgZnJvbSBsYXJnZXIgZGF0YSBiYXNlcyB3aGVuIHRoZSBpbmZvcm1h dGlvbiBiZWNvbWVzIGF2YWlsYWJsZS4NClJlYWwgdGltZSBncmFwaGljcyB3b3Vs ZCBiZSB1c2VmdWwgZm9yIGR5bmFtaWMgdmlld2luZy4NCg0KTGFzdGx5LCB0aGUg YWR2ZW50IG9mIGJvdGggYXVkaW8gYW5kIHZpc3VhbCB0ZWxuZXQgcG9zc2liaWxp dGllcw0KbWVhbnMgd2UgY2FuIGFsc28gb3JnYW5pc2UgaW50ZXJuYXRpb25hbCBk aXNjdXNzaW9uIGdyb3VwcyBvbiBhDQpwYXJ0aWN1bGFyIHN1YmplY3QgUkFQSURM WS4gTW9zdCBtZWV0aW5ncyB0YWtlIHdlbGwgb3ZlciBhIHllYXINCnRvIG9yZ2Fu aXplLCBzb21lIHJlY2VudCB3b3JrIGlzIG1vcmUgaW50ZXJlc3RpbmcgdGhhbiB0 aGF0LiBJbiBhZGRpdGlvbg0Kc3VjaCB0ZWxuZXQgY29uZmVyZW5jaW5nIHdvdWxk IGJyZWFrIHRoZSBpbmZvcm1hdGlvbmFsIG1vbm9wb2xpemF0aW9uDQp3aGljaCBj YW4gb2NjdXIgaW4gbW9yZSBjbGFzc2ljYWwgbWVldGluZy4gIFlvdW5nZXIgc2Np ZW50aXN0cw0Kd291bGQgaGF2ZSBhbiBlYXNpZXIgdGltZSBvZiBnZXR0aW5nIGhl YXJkIGFuZCBpdCB3b3VsZCBhZGQgdG8gdGhlaXINCm93biBtYXR1cmF0aW9uLiBJ ZiB0aGVpciB3b3JrIGlzIGp1bmsgdGhleSdsbCBoZWFyIGFib3V0IGl0IGVhcmx5 DQplbm91Z2ggdG8gY2hhbmdlIGRpcmVjdGlvbnMuIElmIHRoZWlyIHdvcmsgaXMg Z29vZCwgdGhleSdsbCBnZXQgaGVhcmQNCmFuZCBpdCB3aWxsIGNoYW5nZSBvdGhl ciBwZW9wbGVzIHRoaW5raW5nIGFib3V0IGFuIGFyZWEuDQoNCkFsbCB0aGVzZSBh Ym92ZSBpZGVhcyBhcmUgbm90IG15IG93biwgdGhleSB3ZXJlIGV4cHJlc3NlZCBp biBvbmUgZm9ybQ0Kb3IgYW5vdGhlciBieSBvdGhlciBwZW9wbGUuIFNjaWVudGlm aWMgZWRpdG9ycyBjb3VsZCBtYWtlIG1vbmV5IG9uDQpzdWNoIGFuIGUtam91cm5h bHMsIGJ5IG1hcmtldGluZyBhdCBhIGZlZSB0aGUgZGlzdHJpYnV0aW9uIG9mIHRo ZQ0Kam91cm5hbCBhbmQgbWF5YmUgdGhlIHNvZnR3YXJlIG5lY2Vzc2FyeSB0byBj b25maWd1cmF0aW5nIHRoZQ0Kam91cm5hbC4gIFBlcmhhcHMgdGhlIHRpbWUgaGFz IGNvbWUgZm9yIGUtam91cm5hbHMsIGN1cnJlbnQgZS1tYWlsDQpzZXJ2aWNlcyBo YXZlIHRvbyBtdWNoIGVudHJvcHkgYW5kIG5lZWQgdG8gYmUgY29udHJvbGxlZC4g V2hldGhlcg0KZS1qb3VybmFscyB3aWxsIGltcHJvdmUgdGhlIHF1YWxpdHkgb2Yg YXJ0aWNsZXMgSSBkbyBub3Qga25vdywNCmJ1dCBpdCBpcyB2ZXJ5IGxpa2VseSBh biBpZGVhIHdob3NlIHRpbWUgaXMgYXJyaXZpbmcuIEkgYW0gcGxhbnRpbmcNCmEg dGhvdWdodCB0aGF0IHdhcyBwbGFudGVkIGluIG1lIGFuZCB3aWxsIGxldCB5b3Ug Y2FycnkgaXQgZnVydGhlcg0KaWYgaXQgaXMgYW55IGdvb2QuIFRoaW5rIGFib3V0 IGl0IGluIGFueSBjYXNlLg0KDQo= From mwd@carina.cray.com Thu Jun 10 05:13:41 1993 From: mwd@carina.cray.com (Mark Dalton) Message-Id: <9306101513.AA21369@calamity> Subject: Re: DNA secondary Structure To: chemistry@ccl.net Date: Thu, 10 Jun 93 10:13:41 CDT If any one wants I can send you a molecular biology ftp list There is also a e-mail server list. Here is the info for RNA strucuture prediction and display. Mfold (RNAfold) -SGI,Sun,CRAY(I am bias), Dec, VMS Michael Zuker wagner.cbs.umn.edu, frodo.mgh.harvard.edu, a.psc.edu Many other sites, with revisions. Also has multiple suboptimal energies and comparison of the structures from the suboptimal energies to see regions conserved. LoopViewer - MacIntosh ftp.bio.indiana.edu (129.79.224.25) ftp.bchs.uh.edu (129.7.2.43) ViennaRNA - Unix(Sun) - See below Sigfold - Cray, Vax,SGI,Sun,unix - See below a.psc.edu newtree - RNAFOLD phylogenetic comparison software a.psc.edu Dr. S. Le RNA phylogenetic comparison Bruce Shapiro ViennaRNA-1.02: prompt> ftp ftp.itc.univie.ac.at Name (ftp:you): anonymous Password: your_email_adress ftp> cd pub/RNA ftp> bin ftp> get ViennaRNA-1.0.tar.Z ftp> bye This is Version 1.0 of the Vienna RNA Package. ---------------------------------------------- It consists of a few stand alone programs and a library that you can link your own programs with. The package allows you to - predict minimum free energy secondary structures - calculate the partition function for the ensemble of structures - predict melting curves - search for sequences folding into a given structure - compare secondary structures including pairwise alignment. There is also a set of programs for analysing sequence and distance data using split decomposition, statistical geometry, and cluster methods. This part of the package has not yet been tested extensively. The package is known to compile on SUN Sparcs, IBM RS600, HP 720, and Silicon Graphics workstations. Porting to other architectures should present no problems. Installation: To compile this package you need a compiler that understands ANSI C. Check the makefiles in the subdirectories, and edit the CC CFLAGS BINDIR MANDIR and LIBDIR variables to suit your system. Type make all to produce the library and the interactive programs, and make install to copy the executables, library and man pages into the appropriate directories. You may also want to install the header files in ./H in a suitable place. This is not done by the makefile. Also feel free to copy all header files into a single one, if you dislike using a lot of #includes in your programs. There are man pages for all executables, the library is documented in ./man/RNAlib.texinfo. You can either print this file as a manual using TeX (if you have texinfo.tex) or use it as online documentation using emacs. The texinfo.tex macros come with GNU emacs and most other GNU programs. The documentation will hopefully be more complete in future releases. The following executables are provided: RNAfold predict secondary structures RNAeval evaluate energy for given sequence and structure RNAheat calculate melting curves RNAdistance compare secondary structures RNApdist compare ensembles of secondary structures RNAinverse find sequences folding into given structures AnalyseSeqs analyse sequence data AnalyseDists analyse distance matrices All executables read from and write to stdout and use command line switches rather than menus to be easily usable pipes. For more detailed information see the man pages. We have included a patched version of D.G. Gilbert's readseq program for those who often process sequence files from databanks. See the the documentation in that directory for details. If you need help or have any questions send email to Ivo.Hofacker@itc.univie.ac.at Sigfold.readme: INTRODUCTION The package SIGFOLD is used to identify putative RNA functional regions in genomes. It includes three programs of segfold, sigstb and xyplot. The programs segfold and sigstb have different versions which can be executed on CRAY UNICOS, VAX/VMS and IRIS/UNIX. The presentation copy here is a version for IRIS/UNIX. The segfold.f include some subroutines in IMSL library. The xyplot.f is supported by pgplot software. The xyplot subrourine has not been included in segfold and sigstb, however, you can run it after you get x and y data (scores and positions, or scores and windows) by performing segfold or sigstb. Using as a basic premise that some functional regions of RNAs are involved in a distinct folding or an extensive RNA structure, we set out to find those patterns which are significantly different by chance and are highly stable relative to other possible folding segments in the genome. The unusual folding patterns and their locations detected in the sequence are possibly correlated with RNA functional elements. The assumption has been strongly supported by numbers of mutational analyses in Virus and Bacterial Phage systems and extensive computer simulations for these systems. The program segfold and sigstb which predict unusual folding regions and evaluate RNA folding are based on randomized permutations and reconstructions of the observed sequence data. The statistical significance of the RNA folding for a given segment is demonstrated by the significance score, which compares the lowest free energy of the RNA folding of the segment with the mean of the lowest free energies obtained for many randomized sequences having the same base composition as the biological segment sequence. The lower the significance score, the more statistically significant an RNA segment is judged to be in RNA foldings. The thermodynamical stability of a given RNA segment relative to others in the RNA folding which is termed as stability score shows what significant difference there is between the free energy of the folding of the RNA segment and the mean of the minimal free energies computed from all possible RNA segments of the same size in the sequence. Similarly, lower stability scores correspond to a greater likelihood of the folding of an RNA segment relative to other alternative folding of RNA segments in the same sequence. The theoretical analyses for the statistical significance of the stable folding pattern have helped experimental designs for the identification of biological functional elements in RNAs. The program segfold computes the minimum significance and stability scores and their corresponding starting positions of unusual folding segment in the genome for each window from the least window(initial window) to the largest window. As a result the putative functional regions are determined from these score distributions in the sequence. At the same time, it also provides other information such as potentially alternative significant and stable folding segments, as well as distinct open regions which are significantly less stable than randomly shuffled segments and less stable relative to other possible folding regions in the sequence. The program sigstb assesses the statistical significance and relative stability of local RNA segments in RNA folding by scanning the sequence with a given window. The both highly stable and more statistically significant folding region relative to others can be roughly identified by means of the XY-plot of the distribution of the statistical significance and relative stability in the sequence. The unusual folding region can be finely determined by an exhaustive monte carlo simulation (program segfold) in a putative region based on the results derived from sigstb. The program xyplot is a simple program for creating a XY-plot by means of PGPLOT which is a Fortran subroutine package (graphics subroutine library) for drawing simple scientific graphs on various graphics display devices. It includes three modes which are referred to 3 kinds of data files, respectively. PROGRAMS 1. Program Source files and Executable files sigstb.f sigstb segfold.f segfold xyplot.f xyplot 2. Shell Script for Running Program sigstb and xyplot do_segfold.csh do_sigstb.csh 3. Energy Data Files for Programs sigstb and segfold tinoco. energy.table energy.table2 stddev.table stddev.table2 The four *.table files are used to calculate the mean and standard deviation of the lowest free energies of RNA secondary structures from large numbers of randomly shuffled sequences. All these files must be put in the same directory with segfold and sigstb executable files. 4. Test Data Files Sequence file: t4g60.z Job Input files: segfold_test.job sigstb_test.job xyplot_test_1.job xyplot_test_2.job xyplot_test_3.job 5. Output Files Files derived from Running do_segfold.csh and do_sigstb.csh: t4g60_58.min t4g60_100.min t4g60_58.sig t4g60_58.stb xyplot_test_1.plt t4g60_100.sig t4g60_100.stb t4g60_58.open t4g60_100.open t4g60.minsig t4g60.minstb t4g60sig_wd.data t4g60stb_wd.data xyplot_test_2.plt t4g60sigp_wd.data t4g60stbp_wd.data xyplot_test_3.plt 6. Annotations of job files segfold_test_job.note sigfold_test_job.note xyplot_test_job.note -- Mark Dalton AUG-GCU-AGA-AAG H Cray Research, Inc. M A R K | Eagan, MN 55121 CH3-S-CH2-CH2-C-COOH Internet: mwd@cray.com | (612)683-3035 NH2 From cramer@staff.tc.umn.edu Thu Jun 10 05:51:15 1993 Message-Id: <0012c175875012416@staff.tc.umn.edu> From: "Christopher J Cramer-1" Subject: Solvation Models -- Time and Quality To: chemistry@ccl.net Date: Thu, 10 Jun 93 10:51:15 CDT Netters, Andrew Allen has kindly provided an interesting summary of some points made with regard to various solvation models. While comparisons are odious, I'd like to expand slightly (and respond slightly) to a few of these points. > Hi!, > About your question on Amsol, check the following reference: > > "Comparison Methods to Estimate the Free Energy of Solvation: > Importance in the modulation of the affinity of 3-Benzazepines for the > D1 Receptor". I.Alkorta, H.O.Villar, and J.J.Perez > Jour.Comp.Chem Vol.14(5), 620-626 (1993). > > This paper include a detailed comparison between Delphi and Amsol > calculations over a long list of neutral and ionic species. Hope > this helps. > I also recommended this as a non-partisan comparison paper. It's worth noting that Delphi, which solves the Poisson-Boltzmann equation, computes ONLY the electrostatic portion of the free energy of solvation. Moreover, solute electronic polarization is either NOT taken account of (internal dielectric of unity) or is dealt with in a non-quantum way (internal dielectric above unity, typically two). On the flip side, the method is fast, allows variation of solvent ionic strength, and it's trivial to change dielectric. Since the AMSOL models are parameterized to account for non-electrostatic energetic effects of aqueous solvation (e.g. cavitation energy, directional components of hydrogen bonding, etc.), which do not depend in an obvious way on some macroscopic observable which could be obtained for any solvent, it is not YET possible to do other solvents. Another very pretty comparison came from Bill Jorgensen's group, and appeared in either the Feb. or March issue of J. Comp. Chem. (I apologize that my files seem to have eaten my copy, so I can't give the exact reference) this year. This compared Monte Carlo simulations to AM1-SM2 for absolute free energies of aqueous solvation for substituted benzenes. Both methods gave similar mean errors (about 0.6 kcal/mol). While it was pointed out by the authors that AMSOL ran two to three orders of magnitude faster, it was also pointed out that there's a great deal of solvent-structural and thermodynamic information that comes out of the simulation that is not available from the continuum model. Obviously, a choice of methodologies depends at least in part on what you want to learn from your calculation -- it is not a question of one being "better" than another. > > Here at Columbia we have done a lot of work with continuum > solvation models in a molecular mechanics framework. We do very well with > predictions of free energies of solvations for a range of small > organic and bio-organic molecules. We also have success with > prediction of free energies of host-guest binding in solvent and things > like predicting hydrogen bonding populations for small amides in organic > solvents. > > > The main reference for the solvation model is > J. Am. Chem. Soc. 112, 6127, 1990. > The model is called GB/SA [Generalized Born/ Surface Area] and has both terms > which take into account the cavity and Van der Waals terms (proportional > to surface area) and polarization terms. > > Quentin McDonald. > It is important to note that some of the very interesting ideas incorporated into the GB/SA model (which is classical) inspired us to develop the SMx models (which are quantum), most importantly, the idea that many people want to know the WHOLE free energy of solvation -- not just some fraction of it. The key virtue of the quantum model is that it allows for self-consistent solute electronic polarization, where the classical does not. Of course, if this is an effect which will be small, or will cancel out in some comparison you are interested in, than the MUCH faster molecular mechanics approach may give an equivalent answer. > I don't know if AMSOL has been used > on really large systems yet though. > Probably the largest molecule we've done to date is deoxycytidine. Incidentally, the extra time required for solvation becomes a smaller and smaller fraction of the total time as the size of the molecule gets bigger -- the solution of the HF equations becomes the time crunch, as you might expect. Mind you, larger molecules tend to have floppy modes which give rise to numerous low energy conformers, and one can't get around the difficulty of sampling conformational space at the quantum level any more than at the classical level! > > I have been using Amsol 3.0.1 for quite a while. The following experience arises > using DERINU (numerical derivatives turned on, i.e. full geometry optimization > in the AM1 SM2 method) is very time consuming - I left jobs to run for weeks (!) > on the SGI 380, and they would not finish; moreover, there is an apparently > system dependent bug, in the sense that the DUMP routine does not work, i.e. > it does NOT write a restart file every hour (actually it's rewriting function > never worked...). > I'm mystified about Tudor's problem, since we have no trouble with the restart option on any of the machines we've tested (including the SGI 4D/310GTXB and Indigo R4000). > > On the other hand, Ampac 4, distributed by Dr. Andy Holder from Kansas City, has > an improved version of Amsol that runs 50 to 100 times faster (but does not come > with QCPE, you have to pay for it). Hmm, we've only barely fired up our copy of AMPAC 4.0, but it's hard for me to believe in that speed up, since the AMSOL bottlenecks were incorporated without change (i.e. speeding up the gas phase parts couldn't help that much). While not endorsing anyone's commercial product, I can say the following about speed -- The version currently at QCPE is 3.0.3. This solvation portion of the code is essentially exactly what is in AMPAC 4.0. However, the latter program has certainly been improved over AMPAC 2.1, which was modified (and portabilized) for AMSOL, and some speed-up is simply a given, although I haven't got a firm number. Joe Leonard, at Wavefunction, has coded the SMx models into SPARTAN from scratch (i.e. only using our published theory and results, no code) and has shown me essentially identical results to our calculations (whew!) while reporting a speed up over AMSOL 3.0.1 of roughly 5-50. Again, part of this may be in the solvation portion, and part in the solvation-independent framework. Finally, AMSOL 3.5 is to be sent to QCPE in the very near future (less than a month) and affords the user considerable control over speed vs. precision with certain keyword variables. 10-fold speed increases may be had with no loss of precision -- if you're willing to accept 0.2 - 0.4 kcal/mol fluctuations, 20+-fold. Lastly, the effect of geometry relaxation in solution is usually a tiny fraction of the overall solvation free energy (especially for neutral molecules). If you take the gas phase structure and simply do a 1SCF with the solvation model turned on, it runs hundreds to thousands of times faster than a geometry optimization, and gives you roughly the same answer. At least for intial surveys, this is the route to take! I suppose we should be grateful that people complain to us about speed, and simply accept that it's easy to calculate free energy of aqueous solvation. Reminds me of demanding that the deficit be reduced, but complaining about taxes . . . We continue to work on speed issues, with the goal of getting an aqueous calculation down to only twice as long as the gas phase (no time frame for completion, sorry), but state with no contrition that we are more interested in the chemistry than the coding . . . >> Agreement on the fact that AMSOL takes quite a while to do it's solvation >> runs. But then if you really want the results to be good then you need to >> have a little patience. I've been trying to get meaningful data for several >> months using MD, amsol has already given the geometry expected for the system >> i'm interested in, and that's just gas phase. Hopefully the Solvation run >> is going to be even closer to the expected conformation. Always nice to be appreciated, but obviously we can't take credit for this one. The heroic people who parameterized the gas-phase semiempirical Hamiltonian deserve the recognition. I hope people accept this as a balanced review of these issues. While some partisanship is inevitable, it is a pity when it impedes scientific progress. So, I would reiterate that there is no "best" solvation model in general, rather there are specific situations where one or more may be more efficiently applied to a system of interest. Chris -- Christopher J. Cramer University of Minnesota Department of Chemistry 207 Pleasant St. SE Minneapolis, MN 55455-0431 (612) 624-0859 From noy@tci002.uibk.ac.at Thu Jun 10 23:02:15 1993 From: noy@tci002.uibk.ac.at (Teerakiat Kerdcharoen) Message-Id: <9306102002.AA11100@tci002.uibk.ac.at> Subject: Re: Mailing list for Molecular dynamics ? To: chemistry@ccl.net Date: Thu, 10 Jun 1993 22:02:15 +0100 (DFT) > > > Hi netters, > > I am a graduate student in Chemical Eng. at Purdue, > working in Molecular Dynamics Simulations. I would like to know if there > are any mailing lists like this one, which deal primarily with MD simulations, > esp. of polymers. > > Thanks in advance, > > - Samir Kudchadkar > Hallo Samir, I am also working in molecular dynamics simulations and would like to join you. As far as I know, there is no discussion group specifically for this topic. To my best knowledge, there were a lot of discussions about MD some times ago on this mailing lists. Around this moment, I really wonder why people keep silent on this interesting topic. Anyway, I believe that this mailing lists can serve for discussions about MD. Just start to ask to the list. sincerely, Teerakiat ---------------------------------------------------------------------------- NAME : Teerakiat Kerdcharoen (NOY) ADDRESS : Theoretische Chemie Innrain 52 a, A-6020, Austria E-MAIL : noy@tci.uibk.ac.at,noy@tci2.uibk.ac.at,c72454@cx.uibk.ac.at c72454@u1.uibk.ac.at, c70854@tci3.uibk.ac.at noy@atc.atccu.chula.ac.th, noy@atc2.atccu.chula.ac.th ( Thailand ) RESEARCH : Computer Aided Molecular & Material Designs / CAMD ----------------------------------------------------------------------------- From fisher@scripps.edu Thu Jun 10 08:04:46 1993 Date: Thu, 10 Jun 93 15:04:46 PDT From: fisher@scripps.edu (Cindy Fisher) Message-Id: <9306102204.AA05614@struct.Scripps.EDU> To: chemistry@ccl.net Subject: FRODO I believe FRODO is available on SGI in the guise of a program called TOM. We have several versions of it running around here that people use. The original IRIS adaptation was done at UCSD by Stuart Oatley. Cindy Fisher =========================================================================== Cindy Fisher Department of Molecular Biology--MB4 The Scripps Research Institute 10666 North Torrey Pines Road La Jolla, CA 92037 E-mail to: fisher@scripps.edu From raman@bioc01.uthscsa.edu Thu Jun 10 15:54:26 1993 From: raman@bioc01.uthscsa.edu (C.S.RAMAN) Message-Id: <9306110154.AA09743@bioc01.uthscsa.edu> Subject: MOLECULAR DYNAMICS - mailing lists, etc. To: chemistry@ccl.net Date: Thu, 10 Jun 1993 20:54:26 -0500 (CDT) Dear Netters As promised here is the blurb about MD mailing list in addition to information on new MD programs! 1. Some of you must already be familiar with Andreas Windemuth's program MD for doing molecular dynamics simulations on macromolecules. It is extremely well written code (in C) and provides the newcomer with insight into what MD is all about. For those of you who are unaware of this program, you can FTP the same from: lisboa.ks.uiuc.edu The programs are in compressed form (*.Z) in the directory /pub/md/ Users implementing the results obtained from this program in their manuscripts are requested to cite the following references: a) Board, J.A., et al. [1992] Chem.Phys.Lett. 198, 89 b) Windemuth, A. and Schulten, K. [1991] Molecular Simulation 5,353 2. Klaus Schulten's group is also developing a program which will include (in the near future!) capabilities to view the trajectories created by MD. This program is called VRCHEM and is available for FTP from the same site mentioned above! look in the directory /pub/vrchem 3. MD Mailing List! :) There is good bit of interest in a program like MD and as the field develops, there is a need for dissemination of ideas and inprovements! During a recent conversation with Andreas, I indicated to him that there is substantial interest in the Biomolecular Structure community for his program. This, in addition to the Parallelized version of MD (PMD, which Andreas is developing now!) prompted him to setup a mailing list (on June 8, 1993) whose purpose is the following: Distribute the PMD program to be tested by various users and receive feedback; at the same time, let users share their experiences and discuss improvements to the code. NOTE: Andreas does not provide support for PMD at this time. " PMD is a scalable, parallel program for the simulation of the dynamics of biological macromolecules. Its force field is related to programs such as CHARMM, EXPLOR, GROMOS, DISCOVER and others. Emphasis is placed on improved algorithms to significantly enhance efficiency of the simulation of biological macromolecules, without sacrificing long range interactions" I would strongly urge those interested in Molecular Dynamics to subscribe to this mailing-list and contibute to the same. Although, in its present implementation, the list is for questions and discussions related to PMD, I don't see why one could not discuss general md related issues there. In any case, here is how you subscribe to the list: send mail to: pmd-request@cumbnd.bioc.columbia.edu Request that your name be added to the mailing list! That should do... You can also download PMD by FTP'ing to: cumbnd.bioc.columbia.edu the files are located in /pmd/ I hope that this posting (at least transiently!) quenches the thirst for increased discussion in the MD community! Cheers -raman -- C.S.Raman raman@bioc01.uthscsa.edu - Internet UNIX Programming & Administration 70412.2354@compuserve.com - CIS SPARC & SGI Systems raman@hermes.chpc.utexas.edu - CHPC Department of Biochemistry craman@launchpad.unc.edu UTHSCSA 7703 Floyd Curl Dr. (210) 567-6623 [Tel] San Antonio, TX 78284-7760 (210) 567-6595 [Fax] ****************************************************************************** If a man's wit be wandering, let him study the Mathematics -Francis Bacon ****************************************************************************** From raman@bioc01.uthscsa.edu Thu Jun 10 17:52:32 1993 From: raman@bioc01.uthscsa.edu (C.S.RAMAN) Message-Id: <9306110352.AA09821@bioc01.uthscsa.edu> Subject: FRODO To: chemistry@ccl.net Date: Thu, 10 Jun 1993 22:52:32 -0500 (CDT) For those of you interested in TOM (see the posting by Cindy Fisher), my next post will give you sites from which you can FTP the same. -- C.S.Raman raman@bioc01.uthscsa.edu - Internet UNIX Programming & Administration 70412.2354@compuserve.com - CIS SPARC & SGI Systems raman@hermes.chpc.utexas.edu - CHPC Department of Biochemistry craman@launchpad.unc.edu UTHSCSA 7703 Floyd Curl Dr. (210) 567-6623 [Tel] San Antonio, TX 78284-7760 (210) 567-6595 [Fax] ****************************************************************************** If a man's wit be wandering, let him study the Mathematics -Francis Bacon ******************************************************************************