[sharp-discuss] Hendrickson-Lattman coefficients

Tue Feb 19 08:50:04 GMT 2008

Hi Steph,

On Mon, Feb 18, 2008 at 03:23:34PM -0600, Stephanie Swanson wrote:
> Hi Clemens,
> 
> Thanks for responding!
> 
> Unfortunately, I wanted to use external phase info (HLs) to find HA sites.
> I am still pretty new to all of this, so please let me know if my logic is
> incorrect.
> 
> I am plagued with poor data.  I have low resolution data (3.5 angstroms at
> peak wavelength with signal down to 4 angstroms).  I have a possible MR
> soln.  I can make an anomalous difference map and locate 3 peaks that align
> well with methionines in the MR model.

Very good: not only do you get the initial Se positions, you have also
confirmed your MR solution ;-)

> Unfortunately, the peaks are in close proximity to each other (2
> sites are 4 angstroms apart, the third is equidistant from the other
> 2 sites at a distance of 7 angstroms).

Happens quite often with Met - shouldn't be a problem per se.

> I have tried inputting these initial HA sites, but they are not kept
> during refinement.

Not good - but with low resolution data a few outliers or other dodgy
bits of data can throw off the parameter refinement - e.g. leading to
B-factors of the heavy atom of 300.

Another possibility is, that you did the MR in two possible
spacegroups (say P61 and P65): only one gave you that solution. Then
make sure to use the same spacegroup in your autoSHARP run as well:
otherwise the sites will not be correct. So if your MR solution is in
P65 and your MTZ/SCALEPACK file for autoSHARP still has P61 in the
header:

  - define yourself as 'EXPERT' user in the preferences (from the main
    SHARP Control panel)

  - select the P65 spacegroup in the pull-down menu on the second
    autoSHARP input form page

Also: are there any interesting messages from autoSHARP? Any warnings
about data quality, completeness or so?

> I was hoping Hendrickson-Lattman coefficients would help.

In principle yes. but what I would be doing is slightly different:

* your autoSHARP run that started from these 3 sites (found by
  anomalous Fourier using the MR solution) will have created the
  required input files for SHARP. So you should be able to run SHARP
  (and the whole density modification that usually follows it) also by
  hand.

* you can try and include the MR solution into the density
  modification (and also into the SHARP refinement): just copy/link it
  into

    sharpfiles/datafiles/model_SOMETHING.pdb

* say your autoSHARP run was called FirstTrial.0: then your first SHARP
  run is called FirstTrial.1. From the autoSHARP log-file you should
  be able to go to that SHARP logfile

    FirstTrial.1/LIST.html

* at the bottom you can enter the so-called 'Phase Improvement and
  Interpretation Panel' (a long name I know). From there you can now
  do density modification etc.

* run a solvent flattening (optimizing) job for say 50 cycles,
  starting at a resonable initial solvent content (0.5 isn't a bad
  choice). You want to include the MR model for the first say 5-10
  cycles. If you have MSE (i.e. Se-MET residues) in your PDB file:
  specify that it contains the heavy atoms.

  This will hopefully help getting started with a good initial solvent
  mask and therefore a better density modification.

* you can also make use of HLs fro the model _or_ the best density
  modifcation: there are tools for creating an appropriate MTZ file at
  the bottom of the page. I prefer using density modified HLs to feed
  back into SHARP (and include the model during density modification).

I hope that gets you started with some ideas ...

Cheers

Clemens

> 
> Thanks,
> Steph
> 
> 
> 
> On Feb 18, 2008 3:46 AM, Clemens Vonrhein <vonrhein at globalphasing.com>
> wrote:
> 
> > Hi Steph,
> >
> > no - this can only be done in SHARP. But your autoSHARP run will
> > (hopefully) created a sensible starting point for running SHARP -
> > after all, SHARP is the central part of autoSHARP.
> >
> > Or are you trying to use external phase information (in the form of
> > HLs) to find heavy atom sites in the first place?
> >
> > Cheers
> >
> > Clemens
> >
> > On Fri, Feb 15, 2008 at 12:51:51PM -0600, Stephanie Swanson wrote:
> > > Hi,
> > >
> > > I am a fairly new user of SHARP.  I am trying to figure out if there is
> > any
> > > way to input Hendrickson-Lattman coefficients into an autoSHARP run; or
> > can
> > > this only be done in SAHRP?
> > >
> > > Thanks,
> > > Steph
> >
> > > _______________________________________________
> > > sharp-discuss mailing list
> > > sharp-discuss at 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)
> > ***************************************************************
> >

> _______________________________________________
> sharp-discuss mailing list
> sharp-discuss at 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)
***************************************************************