[phenixbb] R/Rfree discrepancy Phenix vs Refmac

David Goldstone david.goldstone at nimr.mrc.ac.uk
Fri Aug 12 06:15:42 PDT 2011

Hi Sabine,

I am by no means an expert in such things and I am sure that you will 
get a comprehensive answer at some point. My understanding is that 
differences in weightings, and the treatment of the solvent correction 
means reproducing R-factors across different programs is not really 

Also in the greater scheme of things they are not too dissimilar.

Kind regards


On 12/08/2011 13:57, Sabine Schneider wrote:
> Hello Partha,
> At the moment I was more looking, if it makes a difference what comes
> out from Refmac, when I feed it different start models; straight from
> Phaser, reset B-values, after shaking coordinates, after shaking
> coordinates and SA, and so on....
> I assumed, if I put the output from Phenix (with R/Rfree X/X reported by
> phenix) to be identical to the R/Rfree Refmac reports as START value
> when it gets directly the output pdb from phenix?
> I thought the reported R/Rfree values are just something like 'model vs
> data'? That's what I am confused about...
> No twinning, no automated water building, just 1mol/asu, no twinning
> Sabine
> On 08/12/2011 02:36 PM, Partha Chakrabarti wrote:
>> Hi Sabine,
>> I believe the comparison should be with same input, not output of one
>> piped into another.
>> How different are the geometry? Different programs use different
>> algorithm and different xray:geometry term by default. Hence, it might
>> only make sense if the RMSD are almost exactly the same. Furthermore,
>> is one program detecting NCS or twining or building waters
>> automatically? Then everything is blown up anyway.
>> In any case, refinemt program A vs. B is an interesting topic! I would
>> not start one.. ;)
>> MFG,
>> Partha
>> On Fri, Aug 12, 2011 at 5:19 PM, Sabine Schneider
>> <sabine.schneider at mytum.de <mailto:sabine.schneider at mytum.de>> wrote:
>>     Hello,
>>     I am refining a structure at 2.1A solved by MR. For curiosity I'm
>>     testing things like 'shaking' coordinates, simulated annealing,
>>     refinement Refmac vs Phenix and so on, to see what influence that
>>     has on stats, maps etc.
>>     For instance after MR I did a bit of shaking the coordinates with
>>     pdbset (noise 0.1), followed by simulated annealing in Phenix.
>>     Phenix states after SA:
>>     Start R-work = 0.2671, R-free = 0.2992
>>     Final R-work = 0.2312, R-free = 0.2666
>>     When I use the output pdb of phenix directly in Refmac (with same
>>     mtz as input for Phenix)
>>     Refmac tells me:
>>     Initial R factor 0.2392 R free 0.2887
>>     So I am quite puzzled about the discrepancy. Or can someone tell
>>     me if I made an error in reasoning somewhere?
>>     Thanks a lot for the help!
>>     Sabine
>>     _______________________________________________
>>     phenixbb mailing list
>>     phenixbb at phenix-online.org <mailto:phenixbb at phenix-online.org>
>>     http://phenix-online.org/mailman/listinfo/phenixbb
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb

David Goldstone, PhD
National Institute for Medical Research
Molecular Structure
The Ridgeway
Mill Hill
London NW7 1AA

More information about the phenixbb mailing list