Hi Tom,
what I meant was that if there is going to be a conformational
change in ligand or receptor after separating them, it may take
the simulation a long time to reach the right structure(s). During that
tims, the energies of the snapshots may be changing, and the
binding free energy will depend on how long you run it. In contrast,
using the same trajectory for the isolated receptor and ligand can
give you a converged value much sooner (though it may not include
the effects of the induced fit, if any).
I'm not an expert in that area, so maybe others can provide more specific
examples or references.
carlos
Thomas Whitington wrote:
>Hi Carlos,
>
>Thankyou very much for your fast reply, it has helped clarify things for me.
>What exactly do you mean by "convergence" of the simulations in this context?
>Does it have something to do with the calculation of a mean and standard
>deviation for delta G, given the multiple snapshots of receptor, ligand, and
>complex?
>
>Cheers,
>Tom Whitington
>
>Quoting Carlos Simmerling <carlos.ilion.bio.sunysb.edu>:
>
>
>
>>In principle, yes. In practice, it has been reported that the
>>data can be more reliable if this is not done. Check some of the
>>published M-PBSA applications; both approaches have been
>>used. The basic issue is that the separate simulations may be poorly
>>converged, and that using the complex trajectory is approximate but
>>may be better converged (particularly if there is a receptor or ligand
>>conformational change on binding). If one wants relative binding
>>energies, the receptor ones cancel anyway.
>>
>>Thomas Whitington wrote:
>>
>>
>>
>>>Hi all,
>>>
>>>I have a question regarding the mm_pbsa examples distributed with amber 8,
>>>
>>>
>>and
>>
>>
>>>in particular, the example of estimating binding free energy.
>>>The general approach for estimating binding free energy is to obtain overall
>>>energy estimates for the recetor, ligand and complex separately, and then to
>>>subtract the ligand and receptor estimates from the complex estimate.
>>>
>>>This approach seems to be taken in the mm_pbsa example (example 3). However,
>>>
>>>
>>the
>>
>>
>>>ligand, receptor, and complex snapshots (generated in example 1) that are
>>>analysed during mm_pbsa energy estimation all come from the same trajectory
>>>
>>>
>>file
>>
>>
>>>(named md_traj_short.mdcrd). In contrast, shouldn't separate "production
>>>
>>>
>>run"
>>
>>
>>>trajectories be generated for the ligand and receptor, after separate
>>>minimisation and equilibration of these entities?
>>>
>>>I have assumed that this approach was taken in the example because it is not
>>>intended to be a thorough example of binding free energy estimation.
>>>
>>>
>>However, if
>>
>>
>>>I have interpreted the example incorrectly I would like to know. Any
>>>
>>>
>>comments
>>
>>
>>>regarding this would be greatly appreciated.
>>>
>>>Kind regards,
>>>Tom Whitington
>>>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Sep 11 2005 - 14:53:00 PDT
