[phenixbb] phenix automr

Randy J. Read rjr27 at cam.ac.uk
Mon Mar 30 01:44:29 PDT 2009


Hi,

Tom has already said most of what I would have said. Most importantly, a 
negative LLG tells you that the model is not fitting the data as well as 
Phaser expects it to, given what you have told it about the model 
completeness and assumed RMS error. It may be that the model is too poor to 
solve the structure, but sometimes there is a clear solution, even though 
the final LLG is negative.

You can only really tell if the solution is a clear one by looking at the 
details of the solution statistics. I would look closely at the log file, 
to see whether the solution has decent signal-to-noise. What you would like 
to see is that the Z-scores (at least for the second copy) are reasonably 
high and, in particular, that the packing check did not eliminate potential 
solutions with much higher Z-scores.

It would probably be worthwhile looking at the section in our web page on 
how to work with difficult cases: 
http://www-structmed.cimr.cam.ac.uk/phaser/documentation/phaser-2.0.html#MR_difficult_cases.

Best wishes,

Randy Read
 
On Mar 30 2009, sbiswas2 at ncsu.edu wrote:

>Hi all,
>
>I tried to use phenix automr by using the gui and also I am using the
>latest version of phenix (courtesy of the phenix development team). I
>asked it to search for two molecules in the asymmetric unit but it finds
>just one also I am getting negative (-29) LLG value for the final
>solution. The map also does not look good. I also asked it to search for
>all possible space group, although I processed the data with P43212 phenix
>automr finds a different one P41212.Will fixing the space group fix the
>problem? Does anyone know what is going on? I tried running phaser
>separately through ccp4 and phaser did find two molecules although the map
>did not look good in this either.
>Shya
>_______________________________________________
>phenixbb mailing list
>phenixbb at phenix-online.org
>http://www.phenix-online.org/mailman/listinfo/phenixbb
>




More information about the phenixbb mailing list