Dear Melanie,
To obtain a reasonable error estimate of R-free you could try to refine your structure against different sets of reflections. You can force the R-free set to be the reflections flagged with 1 or 2, etc. instead of 0 (at least in Refmac) and refine to convergence again. You should get similar (but not the same) R-free values. This will give you a good view of possible R-free values. I admit it is a lot of work (and probably totally over the top) but you can script this quite easily. I tried it a few times and the SD of R-free values is typically a few tenths of a percent(age point).
Cheers,
Robbie Joosten
----------------------------------------
> Date: Tue, 2 Mar 2010 08:44:18 -0800
> From: [log in to unmask]
> Subject: Re: [ccp4bb] Deposition of a BUSTER refined structure
> To: [log in to unmask]
>
> On Tuesday 02 March 2010 07:09:47 Melanie Vollmar wrote:
>> Hi all
>>
>> I did my first deposition of a BUSTER refined structure and I was very
>> surprised when I was asked for an "estimated error of free R value" and
>> "estimated error of bin free R value" as well an "ESD based on Luzzati Plot"
>> during the deposition process.
>
> These instructions are unfortunate, since the use of Luzzati plots
> to estimate error is at this point a historical remnant.
>
> The same 'problem' occurs when depositing structures refined with refmac.
> The error estimates output by the program are based either on the
> maximum likelihood statistics or on Cruickshank's DPI indicator.
>
> Bottom line: Just ignore the request for Luzzati plot esds.
>
> Ethan
>
>
>> I had a detailed look into the refinement log file and I tried the Buster documentation but I have absolutely no clue where I can find these values.
>>
>> Can someone give me a hint where I can find some information about this or where the values are in the log file? If they are not listed anywhere, in which way do they have to be estimated? And what is a Luzzati Plot? I have never heard of such a plot before...
>>
>>
>> Thank you very much.
>>
>> Cheers
>>
>> Melanie
>>
>
>
>
> --
> Ethan A Merritt
> Biomolecular Structure Center
> University of Washington, Seattle 98195-7742
_________________________________________________________________
New Windows 7: Find the right PC for you. Learn more.
http://windows.microsoft.com/shop
|