[gmx-users] Re: Re: T->A mutation for a dimer protein

[Mark Abraham]

On 23/06/2011 7:59 PM, Lishan Yao wrote:
> On 23/06/2011 12:22 PM, Lishan Yao wrote:
>> Dear Chris and Mark,
>>       Thank you both for the response. I did the simulation already and Gromacs only gives me the total dH/dl.
> So what's wrong with dividing by two, like I suggested last time? You've
> got two events and you assert that they're independent, so the
> statistics should be easy.
> ---
> I want an error estimation of the free energy change. not sure how if I only know the sum of the two.

If I say the sum of the heights of n independent people is x with 
estimated error of y, you have useful information about the height of a 
single independent person, and its error estimate. Details vary with the 
details.

Mark

>>    The two mutated sites are more than 20A alway from each other which makes it safe to assume that the interaction between the two sites is small. If I can extract dH/dl for each site, I could get a resonable error estimation of my calculation. Furthermore, if this is reliable I will do different mutations in two monomers (as long as they are away from each other) so that I can save my computational time by 50%. I guess this will involve some code change.
> To which mutation would you assign the energy contribution of the atoms
> around half way between them?
> ---
> I do not intend to assign energy contribution to each surrounding atom.  Gromacs gives dH/dl = dH/dl (monomer 1 site) + dH/dl (monomer 2 site). My question is whether it is possible to extract dH/dl (monomer 1 site) and dH/dl (monomer 2 site) separately. I would think it should be straightforward practically.
>
> Best,
> Lishan
>
>
>
> Mark
>
>> Best,
>> Lishan
>>