[sharp-discuss] Sharp suggestions for 60 Se site structure?

[Clemens Vonrhein]

--RwxaKO075aXzzOz0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Dear Bill,

I doubt if the current way of detecting the sites in autoSHARP (using
direct methods via RANTAN) will be very successful for 60 Se sites. So
SHELXD or SnB are probably better options.

What I would do (nevertheless) is the following:

  1. start autoSHARP and request finding sites, HA refinement and
     density modification.

     - you never know, maybe it does work to find some sites

     - have a very good look at the various data analysis/statistics
       steps at the beginning of each autoSHARP run (especially if
       they contain warnings). Especially helpful are

       * self-rotation and native Patterson

       * outlier analysis

       * R-factors between datasets

       * completeness

       * CC on DANO

       * Matthews parameter analysis

       * f'/f'' and wavelength analysis

     - have a look at the various Patterson maps (iso and ano)
       presented as Harker sections in the detection step. I would
       expect the peak and the FA maps to give the nicest maps.

  2. if this autoSHARP run doesn't work, I would try SHELXD (or SnB)
     to get sites.

  3. use the list of sites from either SHELXD or SnB (as fractional
     coordinates in a file sharpfiles/datafiles/*.hatom) in another
     autoSHARP job. You can select this file - but you also need to
     specify the total number of sites you're looking for (this allows
     to go into autoSHARP with only a subset of sites found
     externally).

     Always request to go till density modification!

  4. How do you know you're on the right track? I found the difference
     in correlation coefficient (on E**2) between the two hands (given
     in the density modification step of autoSHARP) to be a good
     indicator. I would expect a significant difference if the HA
     constellation is correct. With lower resolution data the absolute
     value of CC is probably rather high - but there should still be a
     significant difference.

  5.If you get a (somewhat) significant difference: spend some time in
    analysing the various residual maps of the (probably) correct
    hand. Look for additional sites, disordered Se sites (Met tend to
    do that), wrong f'/f'' values and things like that. Cleaning up
    the residual maps can make your life a lot easier later on.

In general at that resolution some B-factor refinement might not
behave that well. On the other hand, I wouldn't worry about Se atoms
with B > 100 (maybe if they go > 200 I would). If you have the feeling
that HAT_B or SCAL_B are misbehaving: set them to sensible values and
do not refine them.

A question for others: any experience with either low resolution
phasing in SHARP/autoSHARP or with lot's of sites in autoSHARP?

Cheers

Clemens


On Tue, May 27, 2003 at 09:47:51PM +0200, William Shepard wrote:
> Hi,=20
>=20
> I have Se SAD data at the Se K-edge to a meager 4A resolution, do any
> autosharpers suggest a HA & phasing strategy. SnB does not give me any
> convincing results, the sharp runs I have done refine 53/60 sites to high=
=20
> B values (>100A**2).
>=20
> Any comments are welcome. Thanks in advance.
>=20
> Cheers,=20
> Bill
>=20
>=20
> _______________________________________________
> sharp-discuss mailing list
> sharp-discuss@globalphasing.com
> http://www.globalphasing.com/mailman/listinfo/sharp-discuss

--=20

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park     Tel: +44-(0)1223-353033
*  Cambridge CB3 0AX, UK           Fax: +44-(0)1223-366889
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************

--RwxaKO075aXzzOz0
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+1H+D4f+nTPEF25QRAjfHAJ9fhtYrMuQ4yB/L78a/Le29IUWItwCfZ1Yk
rYNj3pUaC+qjmtXIc+rdlCI=
=0eYm
-----END PGP SIGNATURE-----

--RwxaKO075aXzzOz0--