OK you could be right, it's not an issue for me because I never use
the FC/PHIC columns anyway, since it's not clear why you *wouldn't*
want to include the solvent contribution, as opposed to any of the
other contributions to the SF (what's so special about the solvent?).
In fact it's not clear why you wouldn't want to include the D factor
either, since it represents the effect of averaging structures
according the probability of the co-ordinate error, i.e. there's
nothing special about the particular set of co-ordinates that the
refinement happened to converge on, it could with almost equal
probability have converged on any of the structures within ~ 1 sigma
co-ordinate error, so it makes complete sense to average the Fcalcs
for the ensemble by applying the D factor (at least that's what I
understood from Randy!).
It seems to me that if you want to say compute the real-space
statistics (RS-R, RS-CC, RS-RMS-Z or whatever) then the logical choice
is to do it between the 2mFo-DFc and the DFc (incl solvent!) maps,
which is what I always do. If Garib changes the meaning of the FC_ALL
column I will still need some way of getting DFc.
So correction to Hailiang: the FC column doesn't contain D, but you
shouldn't be using it anyway.
Cheers
-- Ian
On Fri, Nov 5, 2010 at 5:00 PM, Garib N Murshudov <[log in to unmask]> wrote:
> Yes. That is true at the moment.
> As soon as I have a little bit time I will change (or add) FC_ALL without D weights also.
>
> Garib
>
> On 5 Nov 2010, at 16:59, James Holton wrote:
>
>>
>> I thought that FC was the unweighted Fcalc, and only FC_ALL included the "D" weights?
>>
>> -James
>>
>> On 11/5/2010 7:22 AM, Ian Tickle wrote:
>>> Refmac doesn't output mFo as such, however it does output DFc: that's
>>> what is in the columns FC& PHIC, or FC_ALL& PHIC_ALL if you want the
>>> total DFc including solvent contribution.
>>>
>>> You could obviously get mFo by multiplying the FP& FOM columns, or
>>> you could do it using the fact that mFo = (2mFo-DFc) - (mFo-DFc), i.e.
>>> subtract the DELFWT/PHDELWT from the FWT/PHWT columns. With the
>>> latter you have to be careful to use the complex structure factors
>>> since PHWT& PHDELWT are not necessarily equal. This will give you
>>> mFo exp(i phi) from which mFo is obtained as the absolute value.
>>>
>>> -- Ian
>>>
>>> On Fri, Nov 5, 2010 at 5:12 AM, Hailiang Zhang<[log in to unmask]> wrote:
>>>> Hi, there,
>>>>
>>>> Is there any way refmac can output the mtzfile including mFo/DFc columes?
>>>>
>>>> Thanks!
>>>>
>>>> Hailiang
>>>>
>>
>
>
|