Re: [gmx-users] on gmx make_ndx
On 6/18/15 1:08 AM, Brett wrote: Dear All, In http://www.gromacs.org/@api/deki/files/198/=gmx-tutorial.pdf, there is make_ndx -f run.gro At the prompt, create a group for GLU22 with “r 22”, ARG137 with “r 137”, and then “q” to save an index file as index.ndx. The distance and number of hydrogen bonds can now be calculated with the commands g_dist -s run.tpr -f run.xtc -n index.ndx -o dist.xvg Here for g_dist, it is the distance between GLU22 and ARG137, does it mean the distance between the center of GLU22 and the cnter of ARG137? If so how does GROMACS determine the center of a residue? If I want to get the distance between one specific atom of GLU22 (for example OE2) and one specific atom of ARG137 (for example NH1), how do I specify the residue and the atom of the residue together in the make_ndx ? g_dist calculates center-of-mass distances. Making index groups of single atoms is easy in text editor (just write the atom number you want) or with make_ndx syntax (type 'help' at the prompt to see examples), e.g. r 22 a OE2 -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] cut-off
Thanks Justin. yes, you are right, my simulation is very sensitive to cut-off. I did not understand what do you meant by '' is the balance of forces correct''? So If I understood you correctly, the cut-off lengths must be chosen from the force filed, yes? Since I am not expert how to choose a force field, could you tell me a proper force field for Benzene all-atom model? Best regards On Thu, Jun 18, 2015 at 1:37 PM, Justin Lemkul jalem...@vt.edu wrote: On 6/18/15 7:25 AM, Faezeh Pousaneh wrote: thanks David. but I am wondering for example if a chosen short cut-off produces correct experimental data, should I still make it larger or PME? Playing with cutoffs to try to force models to behave a certain way. You might get a right answer, but is the balance of forces correct? Force fields are generally tuned to be used with specific nonbonded regimes and shouldn't be altered unless systematic studies of those setups have been done. No that is an all-atom model. Just for pedantic sake, GROMOS96 54A7 is indeed a UA force field, but aromatic H are represented explicitly. So in this case, the electrostatic treatment is indeed important. PME makes the Coulombic cutoff a bit more flexible, but van der Waals interactions are going to be significant in a system like this and can be very sensitive to changes in the cutoff. -Justin Best regards On Wed, Jun 17, 2015 at 2:50 PM, David van der Spoel sp...@xray.bmc.uu.se wrote: On 17/06/15 14:22, Faezeh Pousaneh wrote: Hi, Does the choice of cut-off length depends on the chosen force filed? or it can be chosen such that the simulation produces the experimental values? Cutoff should be used consistently, each force field has their setting, however that does not mean these settings are optimal. Here is a recent paper on organic liquids discussing this. http://pubs.acs.org/doi/abs/10.1021/acs.jctc.5b00190 In my case, I simulation Benzene with Gromos54A7 and I choose 1.2 nm for both LJ and coul interactions, is that fine? Is it united atom? In that case you have no Coulomb. However from our work above it follows that 1.2 nm is too short and for accurate results you need to use PME for both Coulomb and Van der Waals. thanks Best regards -- David van der Spoel, Ph.D., Professor of Biology Dept. of Cell Molec. Biol., Uppsala University. Box 596, 75124 Uppsala, Sweden. Phone: +46184714205. sp...@xray.bmc.uu.sehttp://folding.bmc.uu.se -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] confusion on NPT MD simulation
On 6/18/15 6:28 AM, Ming Tang wrote: Dear Tsjerk and Justin, Thanks for your help. I have tried to use 1000 and 1 to restrained the whole protein during NPT equilibration, but found the protein shape changed as well. Is position restraint compatible with pressure coupling? Restraints are compatible with pressure coupling. Too high of a restraint can generate spurious forces, as Tsjerk mentioned below. Note that restraints are just biasing potentials; they can allow for small changes in the structure. You shouldn't necessarily hike the force constant to some high value to try to absolutely fix the atoms. -Justin Thanks very much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Tsjerk Wassenaar Sent: Wednesday, 17 June 2015 1:53 PM To: Discussion list for GROMACS users Subject: Re: [gmx-users] confusion on NPT MD simulation Hi Ming Tang, You don't need a huge force constant to keep it fixed. Too high forces may give instabilities. And a more gentle one (100-1000) does the trick as well. Cheers, Tsjerk On Jun 17, 2015 4:17 AM, Ming Tang m21.t...@qut.edu.au wrote: Dear Justin, Thank you so much. There are another 2 papers in which the authors said they equilibrated their collagen like this. This really made me confused for quite a long time, and I come to you for help finally. Critical analysis is important. In the steps to perform a simulation page, step 8 indicates that NPT equilibration is for fixing the density. I have been confused about how should I equilibrate my collagen before SMD simulation. I want to keep the shape (do not change its length) of my collagen during equilibration, so that the force-deformation curve will not be changed because of the equilibration process. After knowing that freezing and pressure coupling is incompatible, I tried to use NVT with heavy atoms restrained by using -DPOSRES, and further the simulation to SMD (actually umbrella and direction-periodic are used) directly in NVT. The simulation run smoothly and I got my force-deformation curve, but I am still concerned whether my equilibration is good enough for SMD simulation, and whether the results are credible. If I choose to further equilibrate my collagen in NPT after in NVT, is it feasible to restrain parts (backbone, terminal atoms) of my collagen by adding a huge force constant (say 1*10e6 kJ/mol/nm^2) through -fc in genrestr command, so that its shape will be reserved during the equilibration? Could you please kindly give me some guidance on which is the best equilibration process I show follow to get a good starting point for SMD (umbrella+direction-periodic) simulation? Thank you so much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto: gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Tuesday, 16 June 2015 9:37 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/16/15 5:27 AM, Ming Tang wrote: Dear Justin, I read a paper, in which the author equilibrated the collagen like this: Firstly, a 100 ps NVT MD simulation at a temperature 310 K was performed, in which velocity rescaling algorithm with 1ps coupling constant was used. Rigid bonds were used to constraint covalent bond length, allowing a time step of 2fs. Secondly, a 200ps NPT ensemble with 1bar pressure was used to equilibrate the system while keeping the temperature at 310 K, where Berendsen barostat with 1ps coupling constant was adopted. In this step, the whole protein was held fixed. Lastly, we further equilibrated the system in a NPT ensemble for 400ps, with a more accurate pressure coupling algorithm based on Parrinello-Rahman barostat (Parrinello and Rahman 1981). Only the first and the last C − α atoms of each chain were constrained. I wonder whether they used genrestr to restrain atoms, like utilising -fc to add a quite large force in all directions to them, and considered those atoms are kept fixed during NPT equilibration? People are often lazy (or inconsistent) with language. Holding something fixed might mean huge restraint or it might actually mean frozen (freezegrps in the .mdp). There are functional differences between these two. Atoms are also not constrained (that's a restraint!) in GROMACS, so there may be some loose language here... If you have questions about methods for published work relevant to what you're doing, that's what corresponding authors are for! -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul
Re: [gmx-users] Number of Compute Nodes in Gromacs Expanded Ensemble Simulations
Expanded ensemble doesn't require any special number of nodes (states are visited sequentially, not in parallel) so the rules are the same as for any other free energy simulation. On Thu, Jun 18, 2015 at 4:46 AM, Andreas Mecklenfeld a.mecklenf...@tu-braunschweig.de wrote: Hi, I'm new to expanded ensemble simulations in Gromacs. Is there any rule of thumb for the most efficient number of compute nodes? Thanks in advance Andreas -- M. Sc. Andreas Mecklenfeld Stipendiat Technische Universität Braunschweig Institut für Thermodynamik Hans-Sommer-Straße 5 38106 Braunschweig Deutschland / Germany Tel: +49 (0)531 391-2634 Fax: +49 (0)531 391-7814 http://www.ift-bs.de -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] cut-off
thanks David. but I am wondering for example if a chosen short cut-off produces correct experimental data, should I still make it larger or PME? No that is an all-atom model. Best regards On Wed, Jun 17, 2015 at 2:50 PM, David van der Spoel sp...@xray.bmc.uu.se wrote: On 17/06/15 14:22, Faezeh Pousaneh wrote: Hi, Does the choice of cut-off length depends on the chosen force filed? or it can be chosen such that the simulation produces the experimental values? Cutoff should be used consistently, each force field has their setting, however that does not mean these settings are optimal. Here is a recent paper on organic liquids discussing this. http://pubs.acs.org/doi/abs/10.1021/acs.jctc.5b00190 In my case, I simulation Benzene with Gromos54A7 and I choose 1.2 nm for both LJ and coul interactions, is that fine? Is it united atom? In that case you have no Coulomb. However from our work above it follows that 1.2 nm is too short and for accurate results you need to use PME for both Coulomb and Van der Waals. thanks Best regards -- David van der Spoel, Ph.D., Professor of Biology Dept. of Cell Molec. Biol., Uppsala University. Box 596, 75124 Uppsala, Sweden. Phone: +46184714205. sp...@xray.bmc.uu.sehttp://folding.bmc.uu.se -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Different Cv and Cp
Just for those who may face with similar problem as I had. I found where my mistake was. The problem: I have been obtaining Cv and Cp differently for Lutidine. The obtained heat capacities in Groamcs were consistent with kB T^2 c_P = Var (Enthalpy) kB T^2 c_V = Var (Energy) but not with Cv = (dU/dT)_V Cp = (dH/dT)_P. Answer: The answer was for (dU/dT)_V I had simulated (NVT) at two different temperatures and was obtaining energy derivatives, while the volumes were different at two temperatures. Now I kept the volume the same at two temperature and I get the same results as approach 1. In fact there is a difference in Cv and Cp in my molecule, as it is seen in the link for Benzene: https://books.google.se/books?id=eH_1dIZr-zMCpg=PA171dq=liquid+benzene+Cp-Cvhl=ensa=Xved=0CB8Q6AEwAGoVChMIzoPJ0diWxgIVRI4sCh0wCwl0#v=onepageq=liquid%20benzene%20Cp-Cvf=false Best regards On Thu, Jun 11, 2015 at 8:10 PM, David van der Spoel sp...@xray.bmc.uu.se wrote: On 09/06/15 17:29, Michael Shirts wrote: If the simulation are generating configurations with the Boltzmann probability distribution, the results should the same up to error. Cv and Cp should not be exactly the same, though for liquids at room temperature, they are pretty close (look up the precise numbers for the fluid you are interested in). This is not correct. cP and cV can easily differ 20-30%. Not many cV have been measured but you can compute the difference from other fluctuation formulae, see e.g. Caleman et al. J. Chem. Theory Comput. 2012, 8, 61–74, http://dx.doi.org/10.1021/ct200731v. Note that to get accurate numbers you need quantum corrections as well. My group are working on implementing a method for that in gmx dos. Are you calculating enthalpy as U + PV, where P is the constant APPLIED pressure, not the instantaneous pressure, and PV is in same units? Since the PV term should be quite low for liquids, the two heat capacities should be relatively close (within noise -- fluctuation based calculations I think are noisier). Else you need to check if the Boltzmann distributions are being correctly generated: See the code and paper describing it linked here: https://github.com/shirtsgroup/checkensemble On Tue, Jun 9, 2015 at 11:23 AM, Faezeh Pousaneh fpoosa...@gmail.com wrote: Dear Michael, Can I ask a question concerning your previous email, I followed Cv = (dU/dT)_V Cp = (dH/dT)_P for my lutidine molecule, and I get same values for Cv and Cp. But when I test with kB T^2 c_P = Var (Enthalpy) kB T^2 c_V = Var (Energy) I get 40 J/mol.K difference in Cv and Cp. Mean that fluctuation play big role. Which way of checking I can rely? Best regards On Tue, May 26, 2015 at 3:38 PM, Michael Shirts mrshi...@gmail.com wrote: By definition (more fundamental that fluctuation formulas) Cv = (dU/dT)_V Cp = (dH/dT)_P Run two simulations at different T and estimate the derivatives. On Tue, May 26, 2015 at 5:12 AM, Faezeh Pousaneh fpoosa...@gmail.com wrote: Dear Michael, I still would like to know what was your method you mentioned on last paragraph, just for learning: ''Also, to be sure, you should double check by calculating both heat capacities by finite difference formulas as well with two simulations at T+dt/2 and T-dt/2 -- if the fluctuation and finite difference resutls don't agree within propagated error, then something is off.'' ? thanks Best regards On Mon, May 25, 2015 at 5:10 PM, Faezeh Pousaneh fpoosa...@gmail.com wrote: Dear Andre, thank you for the link, you are probably right, It seems that my molecule has the difference Cp-Cv in the same range as benzene (since it has also ring structure). Best regards On Mon, May 25, 2015 at 4:44 PM, Faezeh Pousaneh fpoosa...@gmail.com wrote: Dear Michael, I use Parrinello-Rahman for barostat and v-rescale for thermostat. Sorry, could you explain more the second paragraph please? I did not get the method. What I checked so far is checking if gromacs correctly gives Cv,Cp= Var(Energy or Enthalpy)/kBT^2 , and I find that it gives. Best regards On Mon, May 25, 2015 at 4:11 PM, Michael Shirts mrshi...@gmail.com wrote: Are you running with the Berendsen thermostat or barostat? The gromacs g_energy functions for heat capacity use the fluctuation formula, and the fluctuations with both of these algorithms are wrong (as should be printed in the log file warning message). Make sure you use ensemble-preserving thermostats if you want fluctuation properties. Also, to be sure, you should double check by calculating both heat capacities by finite difference formulas as well with two simulations at T+dt/2 and T-dt/2 -- if the fluctuation and finite difference resutls don't agree within propagated error, then something is off. On Mon, May 25, 2015 at 5:59 AM, Faezeh Pousaneh fpoosa...@gmail.com wrote: Hi, I do not know why I
[gmx-users] Unit of electric field in gromacs
Dear GROMACS users, I found a conflicting (may be) about Unit of electric field. In gromacs manual at page 8, the Unit is around 10^7 V/m = 10^-2 V/nm, while it is V/nm in output file *_fied.xvg of mdrun. It is a very big difference and makes me confused. Could you please tell me the correct unit of electric field in gromacs? Thank you very much. Bests, Viet Man -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] confusion on NPT MD simulation
Dear Justin, I tried to equilibrate my collagen with all the heavy atoms restrained using define = -DPOSRES in NPT for 300 ps, but found that the collagen became much shorter. The restrained force for the heavy atoms generated through pdb2gmx is 1000 kJ/mol/nm by default, why does the length changed a lot? Thank you. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Thursday, 18 June 2015 9:34 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/18/15 6:28 AM, Ming Tang wrote: Dear Tsjerk and Justin, Thanks for your help. I have tried to use 1000 and 1 to restrained the whole protein during NPT equilibration, but found the protein shape changed as well. Is position restraint compatible with pressure coupling? Restraints are compatible with pressure coupling. Too high of a restraint can generate spurious forces, as Tsjerk mentioned below. Note that restraints are just biasing potentials; they can allow for small changes in the structure. You shouldn't necessarily hike the force constant to some high value to try to absolutely fix the atoms. -Justin Thanks very much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Tsjerk Wassenaar Sent: Wednesday, 17 June 2015 1:53 PM To: Discussion list for GROMACS users Subject: Re: [gmx-users] confusion on NPT MD simulation Hi Ming Tang, You don't need a huge force constant to keep it fixed. Too high forces may give instabilities. And a more gentle one (100-1000) does the trick as well. Cheers, Tsjerk On Jun 17, 2015 4:17 AM, Ming Tang m21.t...@qut.edu.au wrote: Dear Justin, Thank you so much. There are another 2 papers in which the authors said they equilibrated their collagen like this. This really made me confused for quite a long time, and I come to you for help finally. Critical analysis is important. In the steps to perform a simulation page, step 8 indicates that NPT equilibration is for fixing the density. I have been confused about how should I equilibrate my collagen before SMD simulation. I want to keep the shape (do not change its length) of my collagen during equilibration, so that the force-deformation curve will not be changed because of the equilibration process. After knowing that freezing and pressure coupling is incompatible, I tried to use NVT with heavy atoms restrained by using -DPOSRES, and further the simulation to SMD (actually umbrella and direction-periodic are used) directly in NVT. The simulation run smoothly and I got my force-deformation curve, but I am still concerned whether my equilibration is good enough for SMD simulation, and whether the results are credible. If I choose to further equilibrate my collagen in NPT after in NVT, is it feasible to restrain parts (backbone, terminal atoms) of my collagen by adding a huge force constant (say 1*10e6 kJ/mol/nm^2) through -fc in genrestr command, so that its shape will be reserved during the equilibration? Could you please kindly give me some guidance on which is the best equilibration process I show follow to get a good starting point for SMD (umbrella+direction-periodic) simulation? Thank you so much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto: gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Tuesday, 16 June 2015 9:37 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/16/15 5:27 AM, Ming Tang wrote: Dear Justin, I read a paper, in which the author equilibrated the collagen like this: Firstly, a 100 ps NVT MD simulation at a temperature 310 K was performed, in which velocity rescaling algorithm with 1ps coupling constant was used. Rigid bonds were used to constraint covalent bond length, allowing a time step of 2fs. Secondly, a 200ps NPT ensemble with 1bar pressure was used to equilibrate the system while keeping the temperature at 310 K, where Berendsen barostat with 1ps coupling constant was adopted. In this step, the whole protein was held fixed. Lastly, we further equilibrated the system in a NPT ensemble for 400ps, with a more accurate pressure coupling algorithm based on Parrinello-Rahman barostat (Parrinello and Rahman 1981). Only the first and the last C − α atoms of each chain were constrained. I wonder whether they used genrestr to restrain atoms, like utilising -fc to add a quite large force in all directions to them, and considered those atoms are kept fixed during NPT equilibration? People are often lazy (or inconsistent) with language. Holding something fixed might mean huge restraint or it might actually mean frozen (freezegrps in
[gmx-users] energygrps too many, more than 100 energy table
Dear gromacs user: I plan to simulate a system with self defined energy function, it would used about 160 specific energy table. The question is how can I write their long name to energygrps in one line. The follow length is ok, but my energy group is more than that. energygrps = A_A A_C A_D A_E A_F A_G A_H A_I A_K A_L A_M A_N A_P A_Q A_R A_S A_T A_V A_W A_Y B_A B_C B_D B_E B_F B_G B_H B_I B_K B_L B_M B_N B_P B_Q B_R B_S B_T B_V B_W B_Y I hear the length of total energy group name has some limitation, is there any way I can solve this problem. Thanks a lot. Best regards, Zhang -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] energygrps too many, more than 100 energy table
Dear gromacs user: I plan to simulate a system with self defined energy function, it would used about 160 specific energy table. The question is how can I write their long name to energygrps in one line. The follow length is ok, but my energy group is more than that. energygrps = A_A A_C A_D A_E A_F A_G A_H A_I A_K A_L A_M A_N A_P A_Q A_R A_S A_T A_V A_W A_Y B_A B_C B_D B_E B_F B_G B_H B_I B_K B_L B_M B_N B_P B_Q B_R B_S B_T B_V B_W B_Y I hear the length of total energy group name has some limitation, is there any way I can solve this problem. Thanks a lot. Best regards, Zhang -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] confusion on NPT MD simulation
Dear Justin, I tried again, and the position restraint works well in NPT equilibration. Thanks very much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Ming Tang Sent: Friday, 19 June 2015 11:08 AM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation Dear Justin, I tried to equilibrate my collagen with all the heavy atoms restrained using define = -DPOSRES in NPT for 300 ps, but found that the collagen became much shorter. The restrained force for the heavy atoms generated through pdb2gmx is 1000 kJ/mol/nm by default, why does the length changed a lot? Thank you. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Thursday, 18 June 2015 9:34 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/18/15 6:28 AM, Ming Tang wrote: Dear Tsjerk and Justin, Thanks for your help. I have tried to use 1000 and 1 to restrained the whole protein during NPT equilibration, but found the protein shape changed as well. Is position restraint compatible with pressure coupling? Restraints are compatible with pressure coupling. Too high of a restraint can generate spurious forces, as Tsjerk mentioned below. Note that restraints are just biasing potentials; they can allow for small changes in the structure. You shouldn't necessarily hike the force constant to some high value to try to absolutely fix the atoms. -Justin Thanks very much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Tsjerk Wassenaar Sent: Wednesday, 17 June 2015 1:53 PM To: Discussion list for GROMACS users Subject: Re: [gmx-users] confusion on NPT MD simulation Hi Ming Tang, You don't need a huge force constant to keep it fixed. Too high forces may give instabilities. And a more gentle one (100-1000) does the trick as well. Cheers, Tsjerk On Jun 17, 2015 4:17 AM, Ming Tang m21.t...@qut.edu.au wrote: Dear Justin, Thank you so much. There are another 2 papers in which the authors said they equilibrated their collagen like this. This really made me confused for quite a long time, and I come to you for help finally. Critical analysis is important. In the steps to perform a simulation page, step 8 indicates that NPT equilibration is for fixing the density. I have been confused about how should I equilibrate my collagen before SMD simulation. I want to keep the shape (do not change its length) of my collagen during equilibration, so that the force-deformation curve will not be changed because of the equilibration process. After knowing that freezing and pressure coupling is incompatible, I tried to use NVT with heavy atoms restrained by using -DPOSRES, and further the simulation to SMD (actually umbrella and direction-periodic are used) directly in NVT. The simulation run smoothly and I got my force-deformation curve, but I am still concerned whether my equilibration is good enough for SMD simulation, and whether the results are credible. If I choose to further equilibrate my collagen in NPT after in NVT, is it feasible to restrain parts (backbone, terminal atoms) of my collagen by adding a huge force constant (say 1*10e6 kJ/mol/nm^2) through -fc in genrestr command, so that its shape will be reserved during the equilibration? Could you please kindly give me some guidance on which is the best equilibration process I show follow to get a good starting point for SMD (umbrella+direction-periodic) simulation? Thank you so much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto: gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Tuesday, 16 June 2015 9:37 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/16/15 5:27 AM, Ming Tang wrote: Dear Justin, I read a paper, in which the author equilibrated the collagen like this: Firstly, a 100 ps NVT MD simulation at a temperature 310 K was performed, in which velocity rescaling algorithm with 1ps coupling constant was used. Rigid bonds were used to constraint covalent bond length, allowing a time step of 2fs. Secondly, a 200ps NPT ensemble with 1bar pressure was used to equilibrate the system while keeping the temperature at 310 K, where Berendsen barostat with 1ps coupling constant was adopted. In this step, the whole protein was held fixed. Lastly, we further equilibrated the system in a NPT ensemble for 400ps, with a more accurate pressure coupling algorithm based on Parrinello-Rahman barostat (Parrinello and Rahman 1981). Only the first and the last C − α atoms of
[gmx-users] pdb2gmx compiled with MinGW on WindowsPC
Dear GROMACS users: This is a self-response for my query in below. Does anybody uses gromacs (actually 4.6.7) compiled with MinGW on windows PC? I'd been using that without any troubles. However, quite recently, I firstly tried pdb2gmx (gromacs-4.6.7 compiled with MinGW on windows PC) and met the trouble. (I'm not bio-related person and I did not use pdb2gmx ever!) Force field scan does not work with the pdb2gmx with MinGW. I guess it may related to the dirent functions. Any other common programs (e.g., grompp, runmd etc.) runs without troubles. Any info.s on pdb2gmx compiled with MinGW on windows PC will be welcome! It seems no one uses gromacs compiles with MinGW. The problem in above really caused by dirent function on MinGW. MinGW has dirent.h but it does not work for gromacs. Then a solution I found was the followings. In the src/gmxlib/futil.c, change every lines with the structures, #ifdef HAVE_DIRENT_H bra-bra #elif (defined GMX_NATIVE_WINDOWS) bra-bra #endif to #ifdef GMX_NATIVE_WINDOWS bra-bra #elif (defined HAVE_DIRENT_H) bara-bra #endif. Just the change the order to stop using dirent in MinGW. Hope it helps someone. Regards, Makoto Yoneya, Dr. AIST, Tsukuba JAPAN -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] oriientation angle of protein at interface
Hi All, I want to calculate orientation angle of protein at organic/water interface asdone in the following paper:” Replica-ExchangeMolecular Dynamics Simulation of Basic Fibroblast Growth Factor Adsorption onHydroxyapatite” http://pubs.acs.org/doi/abs/10.1021/jp501463r But, I don’t know how todefine interface vector as the molecules at the interface are periodic. Can you please guide me inthis regard Thanks Mahreen -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] CHARMM Pulling LINCS Error
Dear Gromacs Users, I have been trying to pull my protein's center of mass away from the alpha-carbon of a residue. My pull group consists of non-hydrogens at least 3 residues distant from the tether residue, which is restrained. I am using implicit solvent in gromacs 4.5.5, pulling at 0.05nm/ps with a 0.002 dt. I have equilibrated for 100ns, and I do not think that more equilibration is necessary. Equilibration generated no LINCS errors. With pulling and restraint force constants of 3000kJ/(mol*nm^2), I generated LINCS errors at about step 30 of the pull. They were also generated at a pulling rate of 0. I was able to do the same experiment in OPLS-AA with no LINCS errors. I decreased the force constants to 1000, and only got a LINCS error for one of 50 pulls (50 different residues' CAs). Any advice on where I could be going wrong would be greatly appreciated. An example mdp is below. Thanks very much, Todd define = -DPOSRES integrator = sd nsteps = 6000 dt = 0.002 comm-mode = angular nstxout = 5000 nstvout = 5000 nstfout = 5000 nstenergy = 5000 nstlog = 5000 continuation= yes constraint_algorithm = lincs constraints = h-bonds lincs_iter = 1 lincs_order = 4 ns_type = simple nstlist = 0 rlist = 0 rcoulomb= 0 rvdw= 0 coulombtype = cut-off tc-grps = system tau_t = 2 ref_t = 300 ld_seed = -1 pcoupl = no pbc = no gen_vel = no implicit_solvent = GBSA gb_epsilon_solvent = 80 gb_algorithm = Still rgbradii = 0 pull= umbrella pull_geometry = distance pull_dim= Y Y Y pull_start = yes pull_ngroups= 1 pull_group0 = Freeze pull_group1 = Pull pull_rate1 = 0.05 pull_k1 = 1000 pull_nstxout= 50 pull_nstfout= 50 -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] Hessian Matrix Order
Hi, Does some know the exact order of the hessian matrix elements? There are 3N x 3N elements, which match up with the N atoms. But I'm not sure how the x, y, z elements are ordered. I guess it is x1,y1,z1,x2,y2,z2,...xn,yn,zn; but I haven't seen any confirmation of this. It could also be x1,x2,..,xn,y1,y2,...yn,z1,...zn or something else. If someone knows, I would be glad to hear about it. Thanks Eugen -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Gromacs on a Power8 machine
Hi, On Thu, Jun 18, 2015 at 2:46 PM, Baker D.J. d.j.ba...@soton.ac.uk wrote: Hello, I'm not sure that this is the correct place to send my question, however please bear with me and help me to redirect my enquiry if necessary. We recently received delivery of two Power8 machines-one of the machines has a K40 GPU installed. I have read that interesting reports that Gromacs performs well on Power8 systems, especially if there's a GPU card available. I have been doing some experiments, and so far I've have been very disappointed by the performance of Gromacs. Can you please elaborate? Something quantifiable would be useful. Have you compared to some other platform? Which one? How? At this stage I am just working on one of the water examples (1536). I am working with the very latest gromacs (5.1) and I have tried configuring with GMX_SIMD=IBM_VSX, however when I try a parallel run I still get this warnng: WARNING: Using the slow plain C kernels. This should not happen during routine usage on supported platforms. As mentioned on your other gmx-dev post, the non-bonded kernels (which represent most of the i/flops in a run) have not been ported to IBM_VSX yet, but the rest of the code should AFAIK have SIMD support, so without GPUs don't expect great performance. However, with GPU acceleration (where the not yet SIMD ported non-bondeds are offloaded), you should get decent performance. Could someone please advise how I should configure and install Gromacs to get the best out of our Power8 machines. Notes: - You'll likely need at least a K40 @875MHz per socket (for the high-end 12C Power8 CPUs); - The balance of ranks to threads and threads/core will need tweaking; somewhere around 4 threads/socket and 8-12 threads per rank were AFAIR optimal. Don't even try to run one rank with 96 threads/socket (it won't even work :). - I recommend using the FFTW version form github (https://github.com/FFTW/fftw3) where Erik Lindahl has recently added VSX implementation (among others). Cheers, -- Szilárd Best regards, David. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] Number of Compute Nodes in Gromacs Expanded Ensemble Simulations
Hi, I'm new to expanded ensemble simulations in Gromacs. Is there any rule of thumb for the most efficient number of compute nodes? Thanks in advance Andreas -- M. Sc. Andreas Mecklenfeld Stipendiat Technische Universität Braunschweig Institut für Thermodynamik Hans-Sommer-Straße 5 38106 Braunschweig Deutschland / Germany Tel: +49 (0)531 391-2634 Fax: +49 (0)531 391-7814 http://www.ift-bs.de -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] confusion on NPT MD simulation
Dear Tsjerk and Justin, Thanks for your help. I have tried to use 1000 and 1 to restrained the whole protein during NPT equilibration, but found the protein shape changed as well. Is position restraint compatible with pressure coupling? Thanks very much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto:gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Tsjerk Wassenaar Sent: Wednesday, 17 June 2015 1:53 PM To: Discussion list for GROMACS users Subject: Re: [gmx-users] confusion on NPT MD simulation Hi Ming Tang, You don't need a huge force constant to keep it fixed. Too high forces may give instabilities. And a more gentle one (100-1000) does the trick as well. Cheers, Tsjerk On Jun 17, 2015 4:17 AM, Ming Tang m21.t...@qut.edu.au wrote: Dear Justin, Thank you so much. There are another 2 papers in which the authors said they equilibrated their collagen like this. This really made me confused for quite a long time, and I come to you for help finally. Critical analysis is important. In the steps to perform a simulation page, step 8 indicates that NPT equilibration is for fixing the density. I have been confused about how should I equilibrate my collagen before SMD simulation. I want to keep the shape (do not change its length) of my collagen during equilibration, so that the force-deformation curve will not be changed because of the equilibration process. After knowing that freezing and pressure coupling is incompatible, I tried to use NVT with heavy atoms restrained by using -DPOSRES, and further the simulation to SMD (actually umbrella and direction-periodic are used) directly in NVT. The simulation run smoothly and I got my force-deformation curve, but I am still concerned whether my equilibration is good enough for SMD simulation, and whether the results are credible. If I choose to further equilibrate my collagen in NPT after in NVT, is it feasible to restrain parts (backbone, terminal atoms) of my collagen by adding a huge force constant (say 1*10e6 kJ/mol/nm^2) through -fc in genrestr command, so that its shape will be reserved during the equilibration? Could you please kindly give me some guidance on which is the best equilibration process I show follow to get a good starting point for SMD (umbrella+direction-periodic) simulation? Thank you so much. -Original Message- From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [mailto: gromacs.org_gmx-users-boun...@maillist.sys.kth.se] On Behalf Of Justin Lemkul Sent: Tuesday, 16 June 2015 9:37 PM To: gmx-us...@gromacs.org Subject: Re: [gmx-users] confusion on NPT MD simulation On 6/16/15 5:27 AM, Ming Tang wrote: Dear Justin, I read a paper, in which the author equilibrated the collagen like this: Firstly, a 100 ps NVT MD simulation at a temperature 310 K was performed, in which velocity rescaling algorithm with 1ps coupling constant was used. Rigid bonds were used to constraint covalent bond length, allowing a time step of 2fs. Secondly, a 200ps NPT ensemble with 1bar pressure was used to equilibrate the system while keeping the temperature at 310 K, where Berendsen barostat with 1ps coupling constant was adopted. In this step, the whole protein was held fixed. Lastly, we further equilibrated the system in a NPT ensemble for 400ps, with a more accurate pressure coupling algorithm based on Parrinello-Rahman barostat (Parrinello and Rahman 1981). Only the first and the last C − α atoms of each chain were constrained. I wonder whether they used genrestr to restrain atoms, like utilising -fc to add a quite large force in all directions to them, and considered those atoms are kept fixed during NPT equilibration? People are often lazy (or inconsistent) with language. Holding something fixed might mean huge restraint or it might actually mean frozen (freezegrps in the .mdp). There are functional differences between these two. Atoms are also not constrained (that's a restraint!) in GROMACS, so there may be some loose language here... If you have questions about methods for published work relevant to what you're doing, that's what corresponding authors are for! -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests
