Envelope Phasing.

From: David Briggs
Date: 12 March 2012 20:10

Hi CCP4bb,

I would like to ask about "envelope phasing" - specifically with SAXS data.

There are papers (1) and tutorials (2) describing how this might be
done, but I have also found comments on the ccp4bb, such as this one
(http://www.proteincrystallography.org/ccp4bb/message11690.html) which
are somewhat less optimistic.

I get the impression from my reading around that SAXS envelope phasing
is somewhat difficult to do unless you have some NCS you can use to
help the phase extension process. Does anybody have any
opinions/evidence/examples/anecdotes/tips about how SAXS envelope
phasing can be done successfully?

Cheers,

Dave

(1) - eg - http://scripts.iucr.org/cgi-bin/paper?dz5081
(2) - eg - http://www.phaser.cimr.cam.ac.uk/index.php/Using_Electron_Density_as_a_Model

============================
David C. Briggs PhD

----------
From: Anastassis Perrakis


Hi David -

I think you have four hurdles to overcome:

1. How good is a SAXS envelope
2. How will you place it in the right place
3. How will you extend from ~15-20 A to around 4 A
4. How will you extend from 4 A to beyond

Steps 2 and 4 might be the easiest - albeit far from trivial.
Point 1 can be argued...

However, I have not seen anyone cracking 3 in a convincing manner...
(in the absence of serious NCS)

I hope the next emails prove me wrong though!

A.

----------
From: Bosch, Juergen <

Point 3:
Use the saxs model for preliminary phasing of your HA derivative

Jürgen

......................
Jürgen Bosch

----------
From: David Briggs


Hi Francis,

As you may have guessed - my question is stems from my recent
acquisition of some native data for a protein that I have lots of SAXS
data for.

I suspect that more conventional MR is unlikely to work in this case
as I only have a half-decent phasing model for ~30% of the protein (2
copies in the ASU, so 2x15%).

I figured that seeing as I had all this SAXS data (and I have many
datasets collected at multiple synchrotrons that all give very
consistent results), I might as well have a pop at phasing with it. I
have an Fsearch run running at the moment using Flms generated from
the Crysol program (ATSAS package) - no output as yet - does this
program only output results to the log file at the end of the run?

I have about 6 weeks until my next available synchrotron time, where
derivatives will be screened, better resolution native data will
hopefully be obtained, etc. In the meantime...

Cheers,

Dave

----------
From: Alexandre OURJOUMTSEV

Dear David,

I would remind that the "molecular-replacement positionning" of molecular envelopes and some methodological features (in comparison with a conventional MR) of this have been discussed in :

Urzhumtsev & Podjarny (1995). "On the Solution of the Molecular Replacement Problem at Very Low Resolution: Application to Large Complexes". Acta Cryst. D51, 888-895.

Concerning the practical results, at my knowlegde, this was the first step in the ribosome phasing by Nenad Ban in 1999 (while the envelope was from EM and not from SAXS, but this shall not make a difference).

Best regards,

Sacha Urzhumtsev

----------
From: Petr Leiman

Using an envelope or a mask to start phasing even with high NCS is still worse than using a low res model from cryoEM that is characterized by an (almost) proper density distribution.

One has to measure the intensities of the very low res reflections (0 0 1 and the likes) precisely.

The presence of high NCS does not guarantee a success even if your initial model is "good" as described in Section 9 here:
http://www.ncbi.nlm.nih.gov/pubmed/11526317

Petr
------------------
Petr Leiman