Re: [ccp4bb] structural homologs as cross seeds

[Patrick Shaw Stewart]

There are many examples where "cross-seeding" with homologous proteins has
worked, both in classical and recent work.  You should add seed-stocks from
any crystals with significant homology to your routine screening
experiments in the first place.

The important thing is to use *random *screens at first.  This has the
effect of giving you new leads and also optimizing the ones that you have
(plus you can control the number of crystals per drop).

See http://hamptonresearch.com/documents/ramc/RAMC2011_T11_Obmolova.pdf
(Omolova and colleagues have examples where complexes were seeded with
crystals of one of the components and vice versa.)

The excellent review by Stura and co covers most of the theoretical points.
 (1999). Epitaxial jumps. *J. crystal growth*, *196*(2), 250-260.

For practical details see the original paper by D'Arcy (Acta Cryst D:
63.4 (2007):
550-554) and also our paper "Random microseeding: a theoretical and
practical exploration  " *Crystal Growth & Design* 11.8 (2011):
3432-3441.

also http://www.douglas.co.uk/MMS_proc.htm

Acoot, this is also the approach to use with your "cubic" crystals.  Don't
think about it too much, just try it!

The only thing that I can't explain is why this random seeding method,
including cross-seeding, is not more well-known.

Hope it works

Patrick



On 27 December 2013 21:03, Mark van Raaij  wrote:

> the differences are likely to be on the protein surface, and these would
> make the crystal contacts, i.e. it would be unlikely the protein could
> crystallise in the same crystal habit.
> Never say never in crystallisation (i.e. try anything), but I would go for
> other things first like additives, different crystallisation techniques,
> temperatures, concentrations etc.
>
> On 27 Dec 2013, at 20:29, Mahesh Lingaraju wrote:
>
> > Dear all,
> >
> > I was wondering if it sounds logical to use the crystals from a possible
> structural homolog as seeds to induce nucleation ? (in terms of overall
> sequence, the proteins are considerably different but based on sequence
> alignment and structures from other related proteins, it is highly likely
> the protein would have the same structure.)
> >
> > Please comment if any of you had experience with this.
> >
> > Thank you
> >
> > Happy holidays :)
> >
> > Mahesh
>



-- 
 patr...@douglas.co.ukDouglas Instruments Ltd.
 Douglas House, East Garston, Hungerford, Berkshire, RG17 7HD, UK
 Directors: Peter Baldock, Patrick Shaw Stewart

 http://www.douglas.co.uk
 Tel: 44 (0) 148-864-9090US toll-free 1-877-225-2034
 Regd. England 2177994, VAT Reg. GB 480 7371 36

Re: [ccp4bb] LSQKAB

[Paul Emsley]

On 30/12/13 04:10, Doeke Hekstra wrote:

Hi All,

I would like to superimpose a query chain (moving frame) from a rather large 
PDB file (~50 chains) to a target frame specified by a chain in a reference PDB 
file.

  echo "FIT RESIDUE CA 2 TO 63 CHAIN $CHAIN" >> lsqkab.conf
  echo "MATCH RESIDUE CA 2 TO 63 CHAIN A" >> lsqkab.conf
  echo "OUTPUT XYZ" >> lsqkab.conf
  echo "end" >> lsqkab.conf
  lsqkab XYZINM ${f} XYZINF "1AHO.pdb" XYZout ${fileout} 

This doesn't answer your question, but if you have Coot you can use the 
attached script as a work-around (you will have to do some editing).


Paul.



(let ((imol-1 (read-pdb "reference.pdb"))
  (imol-2 (read-pdb "moving.pdb"))
  (c-ids (list "A" "B" "C" ))) ;; or perhaps (chain-ids imol-2)
  
  (for-each (lambda (chain-id)
	  (let ((imol-moving-copy (copy-molecule imol-2)))
		(clear-lsq-matches)
		(add-lsq-match 2 63 "A" 2 63 chain-id 0)
		(apply-lsq-matches imol-1 imol-moving-copy)
		(let*  ((selection (string-append "//" chain-id))
			(frag-mol (new-molecule-by-atom-selection imol-moving-copy selection))
			(frag-mol-file-name (string-append "fragment-" chain-id ".pdb")))
			   
		  (write-pdb-file frag-mol frag-mol-file-name
	c-ids))