From raman@bioc01.uthscsa.edu Thu Oct 28 19:45:26 1993 Received: from thorin.uthscsa.edu for raman@bioc01.uthscsa.edu by www.ccl.net (8.6.1/930601.1506) id TAA29020; Thu, 28 Oct 1993 19:03:58 -0400 Received: from bioc01.uthscsa.edu by thorin.uthscsa.edu with SMTP; Thu, 28 Oct 1993 18:04:57 -0500 (CDT) Received: by bioc01.uthscsa.edu (4.1/SMI-4.1) id AA13167; Thu, 28 Oct 93 18:04:25 CDT From: raman@bioc01.uthscsa.edu (C.S.RAMAN) Message-Id: <9310282304.AA13167@bioc01.uthscsa.edu> Subject: NMR To: chemistry@ccl.net Date: Thu, 28 Oct 1993 18:04:24 -0500 (CDT) X-Mailer: ELM [version 2.4 PL3] Content-Type: text Content-Length: 986 Dear Netters I am trying to determine how many of you that subscribe to CCL are involved with assignment and determination of protein structure, using NMR! Please respond to me, so that we may not waste the bandwidth. Thanks -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 c.raman@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 wtang@rainbow.uchicago.edu Thu Oct 28 20:05:44 1993 Received: from midway.uchicago.edu for wtang@rainbow.uchicago.edu by www.ccl.net (8.6.1/930601.1506) id TAA29094; Thu, 28 Oct 1993 19:25:38 -0400 Received: from rainbow.uchicago.edu by midway.uchicago.edu Thu, 28 Oct 93 18:25:34 CDT Received: by rainbow.uchicago.edu (920330.SGI/920502.SGI) for @midway.uchicago.edu:CHEMISTRY@ccl.net id AA15955; Thu, 28 Oct 93 18:30:56 -0500 Date: Thu, 28 Oct 93 18:30:56 -0500 From: wtang@rainbow.uchicago.edu (Wilfred Tang) Message-Id: <9310282330.AA15955@rainbow.uchicago.edu> To: CHEMISTRY@ccl.net Subject: Time step for Brownian dynamics I'm writing a Brownian dynamics program using the following diffusive algorithm of van Gunsteren and Berendsen: (dt) 1 F(t)-F(t-dt) x(t+dt)= x(t) + ------------ [F(t) + --- ------------ (dt)] + random term (m) (gamma) 2 (dt) ^ | | estimates derivative of F with respect to t dt = time step m = mass gamma = collision frequency F = force x = position I use GROMOS to calculate the force, but I wrote my own simulation algorithm since GROMOS doesn't do Brownian dynamics. I'm interested in studying proteins. I'm wondering what time step dt should be used. Actually, more precisely, what ratio (dt)/[(m)(gamma)] to use. The largest time step I can get out of my program is dt = 0.001 ps, with m = 14 g/mole gamma = 150 ps^-1 giving (dt) ------------ = 5E-7 (ps ps)/(g/mole) (m)(gamma) If I use any larger time step dt, the energy rapidly goes to infinity; to get a good simulation, I probably have to use a smaller time step dt than 0.001 ps. This seems to contradict what's been published in the literature. In particular, looking at a Brownian dynamics simulation on a lipid chain in a membrane bilayer (Pastor, Venable, and Karplus, J. Chem. Phys. 89, 1112 (1988)), the parameters used are: dt = 0.015 ps (much larger than my time step!!!!!!!!!!!!) m = 14 g/mole (I think) gamma = 150 ps^-1 giving (dt) --------------- = 7E-6 (ps ps)/(g/mole) (m)(gamma) I don't constrain my bond lengths, while Pastor et al. use SHAKE to constrain bond lengths. The rule of thumb I've heard is that SHAKE allows a 3 times larger time step, but that still doesn't explain the discrepancy between the time step Pastor et al. used and the time step I can obtain. I'd appreciate it if someone with experience in doing Brownian dynamics simulations can give me some hints and/or explanations. We don't have a lot of experience doing Brownian dynamics simulations at U. of Chicago. Thanks, Wilfred Tang (wtang@rainbow.uchicago.edu)