[gmx-users] Perturbing Proteins
Hi all, I was wondering if there is a principled way to perturb molecules and then see its energy converge to the stationary state via the simulation. Perturbation could mean changing the coordinates of the atoms, changing angles dihedrals etc. Anything related to the topology. I want to see this without the addition of any solvents in the environment. My concern is that if I perturb randomly, the molecule may no more remain in a valid state (bad bond distance or bad angles/dihedrals etc.). Any ideas here? Another question is that if I pick up any molecule from Protein DataBank and run a simulation without any solvent addition, does it converge further or are those molecules already on some minimum energy value attainable? Please correct me if I miss some part of the molecular dynamics pipeline. -- Regards, Sanyam Kapoor Masters in Computer Science Courant Institute, New York University https://www.sanyamkapoor.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] Perturbing Proteins
Thanks Qasim and Joe! @Joe To be more specific, I am in search of valid topologies which are not yet at the energy minima. Since, manually building/generating them would be impractical, I thought if I could take a topology which has achieved its energy minima and do a valid "perturbation", I could then have many such valid topologies which exhibit non-stationarity. On Wed, Apr 4, 2018 at 2:43 AM Joe Jordan wrote: > There are a number of ways to bias a system. Many of them are implemented > in gromacs. You will need to be more specific about what you want here. > > The structures you find in the PDB are in an energy minima but these (lets > call them physical) minima may not be (lets call them theoretical) minima > of the force field you chose. > > On Tue, Apr 3, 2018 at 11:15 PM, Qasim Pars wrote: > > > Hi, > > > > Good question! The nstcomm is the frequency at which the center of mass > > motion is removed. My guess is that using short frequency for the nstcomm > > (e.g. nstcomm=1 while simulation time step is 1 fs, dt=0.001) can > perturbe > > the system. That is, the system may not return to its > starting/unperturbed > > state within a short time interval, while nstcomm is applied at every > step > > on the system. Maybe Justin, Mark or the other developers/users can > comment > > it better than me. > > > > > On 3 Apr 2018, at 23:53, Sanyam Kapoor wrote: > > > > > > Hi all, > > > > > > I was wondering if there is a principled way to perturb molecules and > > then > > > see its energy converge to the stationary state via the simulation. > > > Perturbation could mean changing the coordinates of the atoms, changing > > > angles dihedrals etc. Anything related to the topology. I want to see > > this > > > without the addition of any solvents in the environment. > > > > > > My concern is that if I perturb randomly, the molecule may no more > remain > > > in a valid state (bad bond distance or bad angles/dihedrals etc.). Any > > > ideas here? > > > > > > Another question is that if I pick up any molecule from Protein > DataBank > > > and run a simulation without any solvent addition, does it converge > > further > > > or are those molecules already on some minimum energy value attainable? > > > > > > Please correct me if I miss some part of the molecular dynamics > pipeline. > > > -- > > > Regards, > > > Sanyam Kapoor > > > Masters in Computer Science > > > Courant Institute, New York University > > > https://www.sanyamkapoor.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. > > -- > > 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. > > > > > > -- > Joe Jordan > -- > 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. > -- Regards, Sanyam Kapoor Masters in Computer Science Courant Institute, New York University https://www.sanyamkapoor.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] Perturbing Proteins
Hi Mark, I unfortunately don't have background in MD. Do you mind pointing me to relevant literature? I can take over from there. I believe the key phrases are "alternative configurations of a common topology" and "move sets". I was planning to do MC style simulations but was worried that I might change the topology to an exceptionally bad state. Also, could you point me to places in Gromacs that can help me do this kind of a thing? On Wed, Apr 4, 2018 at 3:42 PM Mark Abraham wrote: > Hi, > > On Wed, Apr 4, 2018 at 9:26 PM Sanyam Kapoor wrote: > > > Thanks Qasim and Joe! > > > > @Joe > > To be more specific, I am in search of valid topologies which are not yet > > > > Generally we refer to those as alternative configurations of a common > topology. > > at the energy minima. Since, manually building/generating them would be > > impractical, I thought if I could take a topology which has achieved its > > energy minima and do a valid "perturbation", I could then have many such > > valid topologies which exhibit non-stationarity. > > > > Monte Carlo style simulations of such systems generate plausible > alternative configurations from many different kinds of move sets (see > literature). Of course many moves would be to a highly non-physical state, > which you would accept only with extremely low probability, and > essentially all of them will lack stationarity (even though there are very > many stationary points, they are only a tiny part of the phase space). > > Mark > > On Wed, Apr 4, 2018 at 2:43 AM Joe Jordan wrote: > > > > > There are a number of ways to bias a system. Many of them are > implemented > > > in gromacs. You will need to be more specific about what you want here. > > > > > > The structures you find in the PDB are in an energy minima but these > > (lets > > > call them physical) minima may not be (lets call them theoretical) > minima > > > of the force field you chose. > > > > > > On Tue, Apr 3, 2018 at 11:15 PM, Qasim Pars > wrote: > > > > > > > Hi, > > > > > > > > Good question! The nstcomm is the frequency at which the center of > mass > > > > motion is removed. My guess is that using short frequency for the > > nstcomm > > > > (e.g. nstcomm=1 while simulation time step is 1 fs, dt=0.001) can > > > perturbe > > > > the system. That is, the system may not return to its > > > starting/unperturbed > > > > state within a short time interval, while nstcomm is applied at every > > > step > > > > on the system. Maybe Justin, Mark or the other developers/users can > > > comment > > > > it better than me. > > > > > > > > > On 3 Apr 2018, at 23:53, Sanyam Kapoor wrote: > > > > > > > > > > Hi all, > > > > > > > > > > I was wondering if there is a principled way to perturb molecules > and > > > > then > > > > > see its energy converge to the stationary state via the simulation. > > > > > Perturbation could mean changing the coordinates of the atoms, > > changing > > > > > angles dihedrals etc. Anything related to the topology. I want to > see > > > > this > > > > > without the addition of any solvents in the environment. > > > > > > > > > > My concern is that if I perturb randomly, the molecule may no more > > > remain > > > > > in a valid state (bad bond distance or bad angles/dihedrals etc.). > > Any > > > > > ideas here? > > > > > > > > > > Another question is that if I pick up any molecule from Protein > > > DataBank > > > > > and run a simulation without any solvent addition, does it converge > > > > further > > > > > or are those molecules already on some minimum energy value > > attainable? > > > > > > > > > > Please correct me if I miss some part of the molecular dynamics > > > pipeline. > > > > > -- > > > > > Regards, > > > > > Sanyam Kapoor > > > > > Masters in Computer Science > > > > > Courant Institute, New York University > > > > > https://www.sanyamkapoor.com > > > > > -- > > > > > Gromacs Users mailing list > > > > > > > > > > * Please search the archive at http://www.gromacs.org/ > > > > Support/Mailing_Lists/GMX-Users_List before po
Re: [gmx-users] Perturbing Proteins
Thank you so much Mark! Let me see what I can do. On Wed, Apr 4, 2018 at 3:57 PM Mark Abraham wrote: > Hi, > > I can only suggest the references from > https://en.wikipedia.org/wiki/Monte_Carlo_molecular_modeling. GROMACS does > not currently have any MC style move set functionality. > > Mark > > On Wed, Apr 4, 2018 at 9:54 PM Sanyam Kapoor wrote: > > > Hi Mark, > > > > I unfortunately don't have background in MD. Do you mind pointing me to > > relevant literature? I can take over from there. I believe the key > phrases > > are "alternative configurations of a common topology" and "move sets". I > > was planning to do MC style simulations but was worried that I might > change > > the topology to an exceptionally bad state. > > > > Also, could you point me to places in Gromacs that can help me do this > kind > > of a thing? > > > > On Wed, Apr 4, 2018 at 3:42 PM Mark Abraham > > wrote: > > > > > Hi, > > > > > > On Wed, Apr 4, 2018 at 9:26 PM Sanyam Kapoor wrote: > > > > > > > Thanks Qasim and Joe! > > > > > > > > @Joe > > > > To be more specific, I am in search of valid topologies which are not > > yet > > > > > > > > > > Generally we refer to those as alternative configurations of a common > > > topology. > > > > > > at the energy minima. Since, manually building/generating them would be > > > > impractical, I thought if I could take a topology which has achieved > > its > > > > energy minima and do a valid "perturbation", I could then have many > > such > > > > valid topologies which exhibit non-stationarity. > > > > > > > > > > Monte Carlo style simulations of such systems generate plausible > > > alternative configurations from many different kinds of move sets (see > > > literature). Of course many moves would be to a highly non-physical > > state, > > > which you would accept only with extremely low probability, and > > > essentially all of them will lack stationarity (even though there are > > very > > > many stationary points, they are only a tiny part of the phase space). > > > > > > Mark > > > > > > On Wed, Apr 4, 2018 at 2:43 AM Joe Jordan > wrote: > > > > > > > > > There are a number of ways to bias a system. Many of them are > > > implemented > > > > > in gromacs. You will need to be more specific about what you want > > here. > > > > > > > > > > The structures you find in the PDB are in an energy minima but > these > > > > (lets > > > > > call them physical) minima may not be (lets call them theoretical) > > > minima > > > > > of the force field you chose. > > > > > > > > > > On Tue, Apr 3, 2018 at 11:15 PM, Qasim Pars > > > wrote: > > > > > > > > > > > Hi, > > > > > > > > > > > > Good question! The nstcomm is the frequency at which the center > of > > > mass > > > > > > motion is removed. My guess is that using short frequency for the > > > > nstcomm > > > > > > (e.g. nstcomm=1 while simulation time step is 1 fs, dt=0.001) can > > > > > perturbe > > > > > > the system. That is, the system may not return to its > > > > > starting/unperturbed > > > > > > state within a short time interval, while nstcomm is applied at > > every > > > > > step > > > > > > on the system. Maybe Justin, Mark or the other developers/users > can > > > > > comment > > > > > > it better than me. > > > > > > > > > > > > > On 3 Apr 2018, at 23:53, Sanyam Kapoor wrote: > > > > > > > > > > > > > > Hi all, > > > > > > > > > > > > > > I was wondering if there is a principled way to perturb > molecules > > > and > > > > > > then > > > > > > > see its energy converge to the stationary state via the > > simulation. > > > > > > > Perturbation could mean changing the coordinates of the atoms, > > > > changing > > > > > > > angles dihedrals etc. Anything related to the topology. I want > to > > > see > > > > > > this > > > > > > &