[sharp-discuss] autoSHARP output - few questions

[Clemens Vonrhein]

Dear Sundar,

some remarks:

 - the latest version of autoSHARP (release November last year) uses
   by default SHELXD for finding sites - if you have SHELXC/D
   installed. It is _much_ better at finding sites than RANTAN. That
   doesn't mean that the RANTAN sites are incorrect though.

 - 2.6A is borderline for ARP/wARP to build into it. So don't be
   surprised that this last step doesn't work.

 - You could use the output of SHARP in other programs for doing
   density modification and/or automatic building. See the explanation
   of the various MTZ files generated at

     http://www.globalphasing.com/sharp/manual/appendix1.html

 - the score for the two hands seem to suggest that one is the correct
   one. And your overall phasing power looks very good.

 - which map are you actually looking at? autoSHARP produces a lot of
   maps: don't even bother looking at the SHARP map (i.e. before
   solvent flattening). You should be looking at the density modified
   map.

   Are you using the interface (httpd) to look at maps? this is
   recommended, since it makes sure that the right columns from the
   MTZ file are used to calculate maps. Otherwise see:

     http://www.globalphasing.com/sharp/manual/appendix1.html#edenflat

 - you don't mention the quality of your data (completeness and
   redundancy are good - but what about I/sig(I), Rmerge etc). Or
   spacegroup (twinning?). Is it anisotropic? What about radiation
   damage - if it is highly redundant, maybe using only the first part
   of your collected data might help.

 - what type of heavy atom and how many residues? From the peak
   heights you report I would guess it is 2 sites only (and the others
   are noise). Try running SHARP (or autoSHARP) only with 2 sites. And
   check the residual maps afterwards visually if additional nice pop
   up nicely.

Hope that gives you some ideas ...

Cheers

Clemens

On Sun, Sep 10, 2006 at 02:19:35PM -0500, Sundaramoorthy, Munirathinam wrote:
> This is a summary of my first ever autoSHARP attempt.  
> 
> I am working with a high redundancy and nearly 100% complete SAD data set of a heavy metal derivative (2.6 A resolution). The number of sites is not known but there are at least two sites as seen from the difference Patterson maps.  Both autoSHARP (RANTAN) and another program give at least 8 identical sites, including the two major ones I could interprete from the Patterson. But the differences in the peak heights between two major sites and others is huge. The first peak is 54 sigma, the second one is 10.5 sigma, and others are between 5-6. The initial CC (test) for 1 peak (> 40 s) is 0.214 and final CC for 5 peaks (> 6 s) is 0.243.  According to the manual,  this falls in ???good??? range. 
> 
>  
> 
> The FOM values  in the first and second SHARP runs are 0.357/0.218 and 0.322/0.197 respectively and the corresponding overall phasing power is 1.925 and 1.65. In the second run an additional site is found but not added as the number of sites exceeds what I asked for.
> 
>  
> 
> If I examine the log file, the phasing power after 3rd Big Cycle drops below 1 around 3.6 A.
> 
>  
> 
> The best and inverted hand score are 0.3894 and 0.3150. The final score is 1.8321 with the suggested solvent content.
> 
>  
> 
> Finally, automatic building gives ???9 chains with 44 residues: 0 docked in sequence [R/Rfree 0.372/0.537] in 10 cycles???. 
> 
> Map doesn???t look interpretable at all.  
> 
>  
> 
> My questions: With the numbers above, does this dataset look promising for SAD phasing? Should I rerun SHARP searching for more sites and cut the resolution to 3.6 A, then do phase modification and extension to 2. 6 A resolution in DM? 
> 
>  
> 
> Further, I have a low redundancy and incomplete native data set at 2.8 A. Would it help or hurt to include this to try SIRAS?
> 
>  
> 
> Thanks in advance.
> 
> 
> Sundar
> 
>  
> 
>  
> 

> _______________________________________________
> sharp-discuss mailing list
> sharp-discuss at www.globalphasing.com
> https://www.globalphasing.com/mailman/listinfo/sharp-discuss


-- 

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park 
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************