Re: [gmx-users] pdb2gmx and periodic molecule
Well, none of us in this project have the scripting skills necessary for this, that's sort of the problem... Thanks, Alex Hi Alex, I think the best way is to extend the chain, such that you get an overlap between both ends. So for a stretch of DNA ACGT You would generate TACGTA Then you strip off the terminal atoms and rewire the links over the boundary. It requires renumbering the topology and at first looks a bit cumbersome, but this ensures all the links are correct and the modifications of the termini have no effect, and it's not very hard to script. And, of course, scripting it is a one time effort to avoid doing it manually later. Cheers, Tsjerk On Apr 26, 2015 5:55 AM, "Alex" nedoma...@gmail.com wrote: JL The force field parameters for a phosphodiester are the same in a linear, finite JL DNA as they will be in your infinite chain. I didn't mean that anything would need to be changed in the description of residues. The FF handles a finite chain perfectly. JL The problem is in telling pdb2gmx JL "don't link i,i+1 here, instead link i,i+(some number of intervening nucleotides)." This is exactly what I meant, the linking. Is there a way to achieve this without messing with the code? I just haven't looked at the code, so if you think I'm being a cheeky little bastard, I'll accept that. :) Thanks, Alex -- Gromacs Users mailing list * Please search the archive athttp://www.gromacs.org/Support/Mailing_Lists/GMX-Users_Listbefore posting! * Can't post? Readhttp://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-usersor send a mail togmx-users-requ...@gromacs.org. -- Best regards, Alex mailto:nedoma...@gmail.com -- 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] pdb2gmx and periodic molecule
Hi Alex, I think the best way is to extend the chain, such that you get an overlap between both ends. So for a stretch of DNA ACGT You would generate TACGTA Then you strip off the terminal atoms and rewire the links over the boundary. It requires renumbering the topology and at first looks a bit cumbersome, but this ensures all the links are correct and the modifications of the termini have no effect, and it's not very hard to script. And, of course, scripting it is a one time effort to avoid doing it manually later. Cheers, Tsjerk On Apr 26, 2015 5:55 AM, Alex nedoma...@gmail.com wrote: JL The force field parameters for a phosphodiester are the same in a linear, finite JL DNA as they will be in your infinite chain. I didn't mean that anything would need to be changed in the description of residues. The FF handles a finite chain perfectly. JL The problem is in telling pdb2gmx JL don't link i,i+1 here, instead link i,i+(some number of intervening nucleotides). This is exactly what I meant, the linking. Is there a way to achieve this without messing with the code? I just haven't looked at the code, so if you think I'm being a cheeky little bastard, I'll accept that. :) Thanks, Alex -- 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] pdb2gmx and periodic molecule
JL The force field parameters for a phosphodiester are the same in a linear, finite JL DNA as they will be in your infinite chain. I didn't mean that anything would need to be changed in the description of residues. The FF handles a finite chain perfectly. JL The problem is in telling pdb2gmx JL don't link i,i+1 here, instead link i,i+(some number of intervening nucleotides). This is exactly what I meant, the linking. Is there a way to achieve this without messing with the code? I just haven't looked at the code, so if you think I'm being a cheeky little bastard, I'll accept that. :) Thanks, Alex -- 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] pdb2gmx and periodic molecule
Justin, I am not entirely sure I understand your sequence of steps. Why do you want to first run it normally, but then also without adding the hydrogens? The hydrogens (a total of two) are only added at the termini by default, and in the final product there's no need for them anyway. Here's what I'm able to understand: run as usual, delete all bonded data relevant to termini, add bonded descriptions for the residues at the boundary, modify pdb to remove termini (actually, I can generate the periodic ones automatically). Am I missing something that should make my life easier? Also, is there a way to automate this by modifying the forcefield? See, right now (on the periodic structure without termini) pdb2gmx sees the first residue and looks it up as one with a terminus, which is really the gist of the error it throws, so I doubt it even gets to the point of looking at the boundary. Is it that the utility behavior cannot be changed at the ff level, and thus our only option is manual modification of the topology? Thank you, Alex JL pdb2gmx isn't designed to intuitively handle such cases. The best way to go JL about it is: JL 1. Run pdb2gmx normally to add H that you might need and to apply patching such JL that you have a phosphate on one end and an alcohol on the other (presumably you JL already have this last part done and have suitable .tdb entries). JL 2. Delete unnecessary atoms from the termini. JL 3. Run pdb2gmx again with -ter (choose None) and -missing so that it doesn't JL complain at you about dangling termini. JL 4. Manually edit the topology to introduce the needed bond, angles, and dihedrals. JL -Justin -- 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] pdb2gmx and periodic molecule
On 4/25/15 11:34 PM, Alex wrote: Justin, I am not entirely sure I understand your sequence of steps. Why do you want to first run it normally, but then also without adding the hydrogens? The hydrogens (a total of two) are only added at the termini by default, and in the final product there's no need for them anyway. I was speaking in generalizations; most starting structures need H added. If that's not relevant, then just skip it. Here's what I'm able to understand: run as usual, delete all bonded data relevant to termini, add bonded descriptions for the residues at the boundary, modify pdb to remove termini (actually, I can generate the periodic ones automatically). Am I missing something that should make my life easier? In theory, if you assign a phosphate as one of your termini (rather than the normal C5'-OH and C3'-OH) then you have to modify very little. Also, is there a way to automate this by modifying the forcefield? See, right now (on the periodic structure without termini) pdb2gmx sees the first residue and looks it up as one with a terminus, which is really the gist of the error it throws, so I doubt it even gets to the point of looking at the boundary. Is it that the utility behavior cannot be changed at the ff level, and thus our only option is manual modification of the topology? It's got nothing to do with the force field and everything to do with software and convention. It's the same case with cyclic peptides or circular DNA. It can be made to work, but that's not really how the software was designed because it's not a common procedure. DNA is assumed to be a linear polymer of finite length. If you don't do that, it's up to you to hack it together. The force field parameters for a phosphodiester are the same in a linear, finite DNA as they will be in your infinite chain. The problem is in telling pdb2gmx don't link i,i+1 here, instead link i,i+(some number of intervening nucleotides). -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] pdb2gmx and periodic molecule
On 4/25/15 12:00 AM, Alex wrote: Ahoy, What I have here is a 100% precisely set up periodic ssdna chain of six residues without termini. The goal is to get the 'polymerization' across the box. So, I boxed the chain and set up the periodicity pretty much perfectly. To test, translation in the direction of periodicity by the periodicity constant results in bonds perfectly recognized by things like VMD or pymol (in the concatenated structure). When running pdb2gmx on the boxed structure, the periodicity seems to be ignored and then expectedly I get a fatal error with a custom FF that works fine on finite chains. x2top sees pbc, but we're not using it here. So, no topology in sight. Any ideas? pdb2gmx isn't designed to intuitively handle such cases. The best way to go about it is: 1. Run pdb2gmx normally to add H that you might need and to apply patching such that you have a phosphate on one end and an alcohol on the other (presumably you already have this last part done and have suitable .tdb entries). 2. Delete unnecessary atoms from the termini. 3. Run pdb2gmx again with -ter (choose None) and -missing so that it doesn't complain at you about dangling termini. 4. Manually edit the topology to introduce the needed bond, angles, and dihedrals. -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] pdb2gmx and periodic molecule
Yeah, I was trying to avoid that last part. You can tell by now how much I dislike uhm doing work. :) I'll try it, thanks. Alex JL 4. Manually edit the topology to introduce the needed bond, angles, and dihedrals. JL -Justin -- 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.