[gmx-users] Re: General query regarding MD simulation

[Dr. Vitaly Chaban]

Dear Dr. Chaban.
>
> Thanks for the insight. I'm not really sure though, how will the choice
> affect my final results. I'm actually trying to look at the changes in the
> protein in terms of backbone and sidechain movements as a result of GDP
> and GTP binding to active site. The presence of Mg ion just complicates
> things.
> My thoughts are like any other residue in protein, for Mg ion ang GDP/GTP
> also, the standard charges should be able to govern the energetics and
> hence the movements in the protein properly. Not sure, how much of a
> difference would the change in charges due to polarization effects of Mg
> ion make to the movements. I feel for my study it shouldn't make much of a
> difference. I may be wrong though. Any thoughts ??
>
> With Regards
> Abhishek Acharya
>
>
Dear Abhishek Acharya -

The rigorous way is to compare binding energies between GDP (or its
chemically relevant fragment) and Mg_2+ derived (1) from MD code; (2) from
one of ab initio methods; HF is not a bad choice here. The comparison
should be done for the same fragment and at T=0K, of course.

If the energy from ab initio appears significantly different, than MANUALLY
adjust a point-charge on Mg, so that the energies become closer. In such a
way, you will not have to introduce changes to the pre-parametrized GDF,
and you will have a trustworthy binding at the same time.


-- 
Dr. Vitaly V. Chaban
MEMPHYS - Center for Biomembrane Physics
Campusvej 55, 5230 Odense M, Kingdom of Denmark