also try MOLREP (CCP4i) - we had a case where AMORE and PHASER did not
work (in our hands), but MOLREP did, using the internal MOLREP
sequence alignment/unequal amino acid trimming feature.
This was a case of a P1 cell with 12 copies of the same molecule, 2.3A
resolution, not very high-quality data, see J Virol. 2008 November;
82(22): 11208–11216.
Mark J. van Raaij
Dpto de Bioquímica, Facultad de Farmacia
Universidad de Santiago
15782 Santiago de Compostela
Spain
http://web.usc.es/~vanraaij/
researcherID: B-3678-2009
On 25 Nov 2009, at 17:42, Roger Rowlett wrote:
I would also try Open-EPMR as well as Phaser. It can sometimes find
solutions that Phaser or Amore cannot, especially for multiple
chains and low resolution. EPMR is especially good at handling high
copy numbers of search models.
Cheers.
Pete Meyer wrote:
A few things to try (or double-check):
1. If you ran phaser with SGALTERNATIVE ALL, make sure the mtz file
you gave to refmac has the same screw axes as the MR solution. If
this is off, it'll lead to higher R-factors.
2. As Fred mentioned, try with a single copy of the protein, and
check the FWT (sigma_a weighted 2Fo-Fc) maps. If the solution is
reasonably correct, you should be able to see some density for the
other components. Also, run some controls. Delete a portion of
your search model before MR, and insure that density for that
region is present in maps phased with the positioned model.
3. Use your seleniums. If your MR solution is correct, then you
should see peaks in model-phased anomalous difference maps. But
confirm this with your form_1 data. In my experience, these are
not very sensitive to model completeness (but that's with a
different protein).
Since you've mentioned you're relatively new at this, you should re-
merge (scala or scalepack) as anomalous for anomalous difference
maps if you haven't done so already (normally I'd assume this was
the case for dealing with anomalous data and not mention it...sorry
if I'm telling you things you already know).
Good luck,
pm
Chao Quan wrote:
Dear CCP4 community:
I am a beginner to crystallography and therefore my apologies if
this question is too simple.
Basically we obtained several crystal forms of the same molecule,
which is a hetero-
trimer containing protein A(18kD), protein B(16kD) and a RNA
segment(40nt or about 15kD).
We have solved the structure of one crystal form(form_1); its
information is as follows:
space group = P 42 22;
unit cell = 126.514 126.514 76.766 90.00 90.00 90.00
resolution = 2.7A;
Rwork = 0.25, Rfree = 0.265;
solvent content = 60%;
Number of molecules per Asymmetric Unit = 1;
Data redundancy = 5;
Data Completeness = 94%;
I am now trying to solve the structure of form_2 crystal using
molecular replacement. So far the information I know about form_2
is as follows:
space group = I 422 or I 41 22;
unit cell = 180.096 180.096 152.530 90.00 90.00 90.00
(unit cell is about 4 times the size of form_1)
resolution = 3.3A; (which is low)
Number of molecules per Asymmetric Unit = 2 or 3;
Data redundancy = 4;
Data Completeness = 92%;
There is no twinning(as shown by Sfcheck);
As shown in"Analyse Data for MR", the first peak is 100 and second
is 68.92; I am not sure if this indicates translation in a
Asymmetric Unit;
The problem is, I can not get a good solution by MR using Phaser
(both I422 and I4122 are tried). When I searched for 3 molecules
per Asymmetric Unit, Phaser did not give solutions at all.
When I used 2/ASU instead, I was able to get some solutions, with
typical statistics as follows:
RFZ=14, TFZ=35.9, PAK=0, LLG=1036; However, these solutions had
high R values(like Rwork=0.59 and Rfree=0.58), which indicated
that they are not solutions at all. Still, I tried refinement
using refmac5, but R values did not go down even after 50 rounds;
sometimes they even increased after refinement.
Besides, the RMS values bond length, bond angle and chiral center
were all 0 as show by refmac5.
I tried limiting resolution range to 15-4A in Phaser, which did
not help either.
Now I am completely stuck. Could anyone give me some advice? I
know this situation is very strange, because I am using the SAME
molecule for MR but can not can a solution.
Thanks a lot,
P.S. 1) Both form_1 and form_2 crystals were grown using
Selenomethionine-containing samples. There are 3 Sel_Met in
protein A and 1 in protein B. 2) A 10-aa internal segment of
protein B is missing in the solved structure, which may indicate
high flexibility.
Chao
--
Roger S. Rowlett
Professor
Department of Chemistry
Colgate University
13 Oak Drive
Hamilton, NY 13346
tel: (315)-228-7245
ofc: (315)-228-7395
fax: (315)-228-7935
email: rrowl...@colgate.edu
