[sharp-discuss] Error during heavy atom finding

[Clemens Vonrhein]

Dear Abhimanyu,

On Tue, Oct 30, 2012 at 05:06:28PM +0100, abhimanyu singh wrote:
> Recently I have collected a Se-SAD data set diffracting at around 1.7A in
> I23 space group and am trying to process it through Autosharp.

Sounds good.

> While running when the program tries to find heavy atom sites using
> ShelxD it terminates reporting "ARRAYS TOO SMALL TO STORE REFLECTION
> DATA - TRY -L1" error.

That seems to point to an error in SHELXD? Which version are you
 running? The latest or the latest beta version of SHELXD?

> Have any one ever encountered such an error and could suggest me any
> possible reason behind it?

You simply seem to have too many reflections (very large cell?). Two
possibilities I could think of:

  1. run autoSHARP with lower-resolution data (eg. cut at 2.5A): that
     should be enough for finding your Se sites (if it is possible at
     all).

     If you run this autoSHARP job only to the heavy-atom detection
     stage, you will get a set of initial sites (in a *.hatom file -
     basically fractional coordinates). Put this *.hatom file into
     your sharpfiles/datafiles directory.

     Then re-run the same autoSHARP job again, but:

       * use all data up to 1.7A

       * pick the *.hatom file you just copied over.

     Now autoSHRP will skip the HA detection with SHELXD (avoiding
     that error) and start with the sites found in the previous 2.5A
     run. But since you have all data now, it should give you phases,
     maps and (hopefully) model to 1.7A.

  2. you could write a shelxd wrapper to supply this -L1 flag (never
     came across that before though) and point
     $SHARP_home/machines/*/shelx.setup to the location with that
     rapper (and a link to the true shelxc binary)?

I would go for option 1: it also shows you how one can run autoSHARP
with a starting set of known HA positions (which could come from all
kind of external programs or calculations you did). This is very
sueful if the default way of running SHELXC/D inside autoSHARP doesn't
give you a good enough answer ... or stops with a warning message like
you just hit upon.

Kind regards

Clemens Vonrhein

-- 

***************************************************************
* 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)
***************************************************************