Re: [ccp4bb]: ARP/wARP phase improvement from MAD data

[Anastassis Perrakis <perrakis@nki.nl>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***


On Tuesday, December 3, 2002, at 02:06 PM, Yu Wai Chen wrote:

> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
>
> Dear all,
>
> This is my first attempt to solve a structure by MAD phasing.  I have 
> collect a set of three wavelengths SeMet data to 2.5A, with which I 
> only used data from 20-3 A to locate 11 (out of 12) Se atoms in the 
> A.U., using SOLVE.
Using 2.5 A data might get you site 12 and better phases ...
>
> I have native data to 1.7A and the crystals are more or less 
> isomorphous.
>
> I would like to try to use ARP/wARP to:
> 1. phase improvement and extension using free-atom models, and then
you dont want to do that really ...thats not the best way to do phase 
extension!
> 2. auto build with warpNtrace
you want that one !! ;-)
but before warpNtrace make sure you get the best possible phases for it 
!
>
> Can you gurus advise me how I should proceed?  I am very confused with 
> which mtz files to use, at different stages.  I now have the solve.mtz 
> (20-3A) with some phases and FOMs in it, and the whole SeM data set 
> (truncate output to 2.5A) and a native data set (truncate output to 
> 1.8A).
>
1. re-run Solve with 2.5 A data (if you know the Se sites, mlphare will 
also do, sharp will also do for sure )

2. if coming from solve run Resolve and it will produce a nice mtz file 
with modified phases and  (i hate to say that ...) a very good pdb file 
with a significant portion of the model - so you can skip arp/warp 
<--that's a joke, please use arp/warp !

3. use 'merge mtz files' from ccp4i to combine the 2.5 reSolve .mtz 
file with your native .mtz 1.7 data

4. use the resulting mtz file for arp/warp. you can now:
	a. either start from the Resolve model
	or
	b. start from the mtz file from resolve phases (the mtz file is there 
before the pdb of resolve ...)

..... at 1.8 A arp/warp will do (I love to say that one ... sorry Tom) a 
better job (i hope that is)  than resolve (although that will be 
cheating since resolve will only know the 2.5 A data) but as far as you 
are concerned you have a good chance to get a nicely refined model.

I would personally be tempted to:

use the native in phasing (nice hint to start a flamewar ?)

or actually

if the native is 'less' than 'more' isomorphous to the MAD dataset I 
would suggest to use DMmulti with phases from the MAD cell and refine a 
unity matrix for the 1.8 data cell. dmmulti will 'transfer' good phases 
to the 1.8 A dataset and these would make arp/warp's life easier.

if native and MAd are really non - isomorphous I would also consider to 
get the Resolve model from the 2.5 A dataset, do a quick MolRep to put 
it at the 1.8 A dataset and run ARP/wARp from 'existing mode' using the 
solution from the molrep.

	

	Tassos


> Thanks for your advises.
>
> Wai
> ===================================================================
> Yu Wai CHEN, Ph.D. ..................   email:ywc@mrc-lmb.cam.ac.uk
> Centre for Protein Engineering,             tel:+44-(0)1223-402148
> MRC Centre, Hills Rd, Cambridge CB2 2QH, UK fax:+44-(0)1223-402140
> WWW homepage: http://www.mrc-cpe.cam.ac.uk/~ywc
>
>