Re: [AMBER] Sufficiently equilibrated or not

From: Brian Radak <brian.radak.gmail.com>
Date: Sun, 21 Feb 2021 11:29:07 -0500

some possibly relevant questions and experiences :

- which RMSD are we talking about? Of the ligand only or the whole complex?
Is this after alignment to the receptor? The Ligand? I'll assume ligand
RMSD in the receptor frame, in which case the magnitude of fluctuations may
still not be as unreasonable as one might think. I'll echo Carlos' point
that visualization is critical here. Specific interactions are also much
more sensitive and useful.

- There is actually very little data to suggest that long equilibration (>
~10-50 ns, say) improves reliability, precision, or accuracy of binding
free energy predictions. In fact there is mostly evidence to the contrary.
I'd accept that this could be "right answer, wrong reason" but the
practical observation remains.

- a slightly more anecdotal observation - for some ligands, even ones that
should be strongly bound, minor force field changes can dramatically impact
the ensemble of poses. I can't offer specific advice here, but I think this
is a huge area for advancement that requires good case studies and careful
observation.

HTH,
BKR

On Sun, Feb 21, 2021, 8:59 AM Carlos Simmerling <carlos.simmerling.gmail.com>
wrote:

> I think it's important to visually examine your trajectories, not just
> calculate rmsd. is that the rmsd of the ligand fit to itself? if so it can
> stay low even if the ligand dissociates. It might be better to calculate
> distances for the key ligand-DNA contacts and make sure they are staying
> close.
>
> for your DNA rmsd, 5A seems quite high, but it really depends strongly on
> the system you have. Is the DNA long? Normal bending motions can give large
> RMSD, as can fraying of ends. These might be fine and are unlikely to
> affect your free energy if they are not close to the ligand.Also, the RMSD
> for the complex does reach a plateau around 4-5A about halfway through. You
> cannot tell, though, if the structure has stabilized because it may still
> be chaning conformation but staying 5A from the reference. I prefer taking
> RMSD to the final structure (or better, to a cluster representative, if you
> can do that analysis). Then use that as the reference structure for RMSD
> and see if the RMSD values are low for the second half with respect to that
> structure.
>
> Overall the issue is that these RMSD plots do not give enough information
> to know if the system is behaving well or not, since we don't know much
> about what you are simulating and how you are calculating the RMSD values.
> I think it's best not to use broad measures like overall RMSD, but to use
> your knowledge of the system to analyze specific properties and see if they
> are reasonable (for example, DNA helical properties, ligand-DNA hydrogen
> bonds, and so on).
>
> On Sat, Feb 20, 2021 at 9:34 PM Delwakkada Liyanage, Senal Dinuka <
> sdd313.msstate.edu> wrote:
>
> > Hello Dr. Foley,
> >
> > Yes, the simulations were done in the NPT ensemble. I looked at the
> energy
> > plots and it seemed to me that the energy convergence is attained. Below
> is
> > the attached snapshot of total energy vs simulation progress plot. Also,
> as
> > Dr. Simmerling suggested I looked in to the rmsd of the ligand, and the
> > ligand was stable throughout the last 100 ns of simulation time (image
> > attached). Nevertheless, I thought of extending my simulation time to
> > another 100 ns due to previous fluctuations in the complex rmsd.
> >
> > [cid:355520f2-ea0b-4298-8278-ed731b561041]
> >
> > [cid:785d16f5-99fc-4d41-b210-bcd7636d70c1]
> >
> > Thank you
> >
> > Senal Liyanage
> > ________________________________
> > From: Lachele Foley <lf.list.gmail.com>
> > Sent: Saturday, February 20, 2021 3:30 PM
> > To: AMBER Mailing List <amber.ambermd.org>
> > Subject: Re: [AMBER] Sufficiently equilibrated or not
> >
> > With RMSD, the issue might be more about convergence than
> > equilibration. But, you're certainly not ready for TI.
> >
> > Which ensemble are you in? Is it NPT? If so, what does your total
> > energy plot vs simulation progress look like?
> >
> > On Sat, Feb 20, 2021 at 2:58 PM Delwakkada Liyanage, Senal Dinuka
> > <sdd313.msstate.edu> wrote:
> > >
> > > Hello Dr. Foley,
> > >
> > > Thank you very much for your valuable input.
> > >
> > > Senal Liyange
> > > ________________________________
> > > From: Lachele Foley <lf.list.gmail.com>
> > > Sent: Saturday, February 20, 2021 1:55 PM
> > > To: AMBER Mailing List <amber.ambermd.org>
> > > Subject: Re: [AMBER] Sufficiently equilibrated or not
> > >
> > > It's due to the rising nature, mostly. But, you also need stable
> > > fluctuations. That is, if you have a significant change in the
> > > amplitude of the fluctuation or if there is a trend (rising or
> > > falling) in the data, then the system is not equlibrated.
> > >
> > > On Sat, Feb 20, 2021 at 2:46 PM Delwakkada Liyanage, Senal Dinuka
> > > <sdd313.msstate.edu> wrote:
> > > >
> > > > Hello Dr. Foley,
> > > >
> > > > Thank you for your feedback. If I may ask, is that due to the rising
> > nature of rmsd or rapid fluctuations of the points?
> > > >
> > > > Senal Liyanage
> > > > ________________________________
> > > > From: Lachele Foley <lf.list.gmail.com>
> > > > Sent: Saturday, February 20, 2021 1:41 PM
> > > > To: AMBER Mailing List <amber.ambermd.org>
> > > > Subject: Re: [AMBER] Sufficiently equilibrated or not
> > > >
> > > > I would not consider the system in that image to be equilibrated.
> > > >
> > > > On Sat, Feb 20, 2021 at 2:17 PM Delwakkada Liyanage, Senal Dinuka
> > > > <sdd313.msstate.edu> wrote:
> > > > >
> > > > > Dear Amber Users and Developers,
> > > > >
> > > > > Here I have attached is a rmsd plot obtained for a ~100 ns
> > equilibrated trajectory (saved at every 10 ps) of a DNA/ligand system. I
> am
> > attempting to calculate absolute binding free energy through a
> > thermodynamic integration (TI) approach. I was hoping if you could
> clarify
> > to me whether the system is equilibrated enough to proceed with the TI
> > calculations. I was under doubt due to the fluctuations shown throughout
> > the last 6000 frames (60 ns). I understand there isn't an exact answer to
> > this. But I appreciate your opinion on this.
> > > > > [cid:f8200194-9003-4107-accf-086dd0f15353]
> > > > >
> > > > > Thank you
> > > > >
> > > > > Senal Liyanage
> > > > > _______________________________________________
> > > > > AMBER mailing list
> > > > > AMBER.ambermd.org
> > > > >
> >
> http://secure-web.cisco.com/116hL3r26c60Pja8rn_aYFWEm7xUTHXaZuKnMzMO7EeTnyrUrIjWOChCXCV4tzsu9b670bcf4lJMFhcJx8DtrtoCKNvzMA9U_1l3OQnAE5ZIbVojX2uNG9PGBLKlm6I7sswRsWwO337AMzdK7070amGhn7CsrHDVUQ3YPBcZAArNxm2mTaKbv3RyFFvKnWiEwDB5nF3HJ3CI79UIRTKh25CT1MfW86LhFw4CuiH0VK3nPYRhUGs8ZiV1P4TdJhMnO9C75M4yctjzmP9fKYnd_lJS-Q4czdX9Rsj4VaS_meUiQLk6GIsLImCMY9dFfwO6e/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> > > >
> > > >
> > > >
> > > > --
> > > > :-) Lachele
> > > > Lachele Foley
> > > > CCRC/UGA
> > > > Athens, GA USA
> > > >
> > > > _______________________________________________
> > > > AMBER mailing list
> > > > AMBER.ambermd.org
> > > >
> >
> http://secure-web.cisco.com/116hL3r26c60Pja8rn_aYFWEm7xUTHXaZuKnMzMO7EeTnyrUrIjWOChCXCV4tzsu9b670bcf4lJMFhcJx8DtrtoCKNvzMA9U_1l3OQnAE5ZIbVojX2uNG9PGBLKlm6I7sswRsWwO337AMzdK7070amGhn7CsrHDVUQ3YPBcZAArNxm2mTaKbv3RyFFvKnWiEwDB5nF3HJ3CI79UIRTKh25CT1MfW86LhFw4CuiH0VK3nPYRhUGs8ZiV1P4TdJhMnO9C75M4yctjzmP9fKYnd_lJS-Q4czdX9Rsj4VaS_meUiQLk6GIsLImCMY9dFfwO6e/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> > > >
> > > > _______________________________________________
> > > > AMBER mailing list
> > > > AMBER.ambermd.org
> > > >
> >
> http://secure-web.cisco.com/1nYcGarqXbx_X59IqL9sLgdETxc9asWQp1ph2TDPhoZNtf8MLpuKstvTAL3tvlduRQFHc6_7TAB038IzCcxgqGOFzbmviUVmN9Xjt_-sx1zlCDnDRw-QVRRam3kes3N2lfV4NUlqAESRpBP7Dm9hmtBPtrFCH3NGddauxNM0PpVwkX_5gnhBSPeF3JICpfuYIXQv_AdP2CeU4Y0IOfhUavyGnFg1mOuI6YyvLP7CaqZIE4b9x1t2ubP3qpFahX_T8NBkXFcZPGTF0m8ZDRlSe7XydLkO0IfCNv5WQaCYq-OnMHNWk8Kg8kqYm_RcP6gVR/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> > >
> > >
> > >
> > > --
> > > :-) Lachele
> > > Lachele Foley
> > > CCRC/UGA
> > > Athens, GA USA
> > >
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > >
> >
> http://secure-web.cisco.com/1nYcGarqXbx_X59IqL9sLgdETxc9asWQp1ph2TDPhoZNtf8MLpuKstvTAL3tvlduRQFHc6_7TAB038IzCcxgqGOFzbmviUVmN9Xjt_-sx1zlCDnDRw-QVRRam3kes3N2lfV4NUlqAESRpBP7Dm9hmtBPtrFCH3NGddauxNM0PpVwkX_5gnhBSPeF3JICpfuYIXQv_AdP2CeU4Y0IOfhUavyGnFg1mOuI6YyvLP7CaqZIE4b9x1t2ubP3qpFahX_T8NBkXFcZPGTF0m8ZDRlSe7XydLkO0IfCNv5WQaCYq-OnMHNWk8Kg8kqYm_RcP6gVR/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> > >
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > >
> >
> http://secure-web.cisco.com/1pH31Gockmoi_e7miSy2KRwpLsRBlztgTj93vqsvHQemb67xixbWlEQfMIKdjxeNu0iq9wlFYcWR6FwI7XRZWRAnPSTLYL17ZBx56Jxkt63SNVS9elmmOOhFLinhCrkwpQOMbfC-HOs8T-DHUv8-xOV5w6v_zvwvCBecjHPHKa4Wu-wOmPr8ZMSdi67dR9vFe3NJjWi8SFAgLrDvOGRJlu-nEXOMYQTgwfcANdCCCacZhrwLaG1GrFHg6qfyCwV0ODZaCbV6Z0ocIimcNw6fXx61CqspvzdiRdpaZvjDeeGkEQ-uyYywr_NjrrT8eF0lF/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> >
> >
> >
> > --
> > :-) Lachele
> > Lachele Foley
> > CCRC/UGA
> > Athens, GA USA
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> >
> >
> http://secure-web.cisco.com/1pH31Gockmoi_e7miSy2KRwpLsRBlztgTj93vqsvHQemb67xixbWlEQfMIKdjxeNu0iq9wlFYcWR6FwI7XRZWRAnPSTLYL17ZBx56Jxkt63SNVS9elmmOOhFLinhCrkwpQOMbfC-HOs8T-DHUv8-xOV5w6v_zvwvCBecjHPHKa4Wu-wOmPr8ZMSdi67dR9vFe3NJjWi8SFAgLrDvOGRJlu-nEXOMYQTgwfcANdCCCacZhrwLaG1GrFHg6qfyCwV0ODZaCbV6Z0ocIimcNw6fXx61CqspvzdiRdpaZvjDeeGkEQ-uyYywr_NjrrT8eF0lF/http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Feb 21 2021 - 08:30:02 PST
Custom Search