[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [ccp4bb]: spacegroup problem
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
David & Sameeta,
This type of disorder has been around for a long time, and was the subject
of a few papers in the 1950's with regard to hemoglobin derivatives.
Bragg & Howells (1954) "X-ray Diffraction by Imidazole Methaemoglobin" Acta
Cryst. 7, 409
Cochran & Howells (1954) " X-ray Diffraction by a Layer Structure Containing
Random Displacements" Acta Cryst. 7, 412
The fact that the indexing is "ambiguous" depending on the spot selection
criteria is telling: don't always rely on an indexing "score" to give you
the "correct" answer. Whenever there is a question regarding indexing,
often the best place to start looking is at the original diffraction images
themselves. There is a wealth of data that is often overlooked
(e.g.-systematic "absences" that aren't really absent, systematically
"present" reflections that are a result of overflows from neighboring
reflections in closely spaced lattices, anisotropy along certain directions
correlating with poor packing, etc.....).
The observation that the reflections corresponding to a P4 lattice (in Dr.
Leys case) are weak, but present, means that this crystal form cannot
physically be I4122! From the description (random, but mutually exclusive,
packing within the same crystal), I would argue that it actually cannot be
assigned to a specific space group (see first paper mentioned above for an
explanation). Granted, this is a seperate issue from the actual structure
determination. For instance, obviously to a first approximation, the
crystals can be assinged to space group I4122. If this approximation is
sufficient to give useful information to address the
biological/chemical/structural question at hand, then great. Otherwise, you
will definitely need to find another crystal form.
Again, don't forget to do a reality check against the raw data. Modern
indexing algorithms are quite clever, but sometimes that cleverness results
in misleading the user who places too much trust in them.
Cheers!
Felix Vajdos
Research Scientist
Pfizer Inc
Eastern Point Road
Groton, CT 06340
860-715-6504
-----Original Message-----
From: Leys, Dr D. [mailto:dl37@leicester.ac.uk]
Sent: Saturday, March 29, 2003 9:06 AM
To: sameeta bilgrami; ccp4bb@dl.ac.uk
Subject: RE: [ccp4bb]: spacegroup problem
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear Sameeta,
we have a troublesome case that sound a lot like yours. We have exactly the
same behaviour for crytals obtained for a protein complex, diffracting to
1.8 A. We can either index in P4122 or I4122 depending on how you play
around with the programs and how long you expose your crystals for (all P4
specific reflections are very weak). We could solve the structure using MIR
(both in P4122 and I4122) leading to great maps, but we could only build 1
out of the 2 proteins.....and that structure refines to approx 27% R 31%
Rfree...... Turns out that partner 1 makes up the entire lattice with
partner 2 only binding to partner 1 without contributing to the lattice.
However, partner 1 is a tetramer that has 1 monomer in the AU, the monomeric
partner 2 binds to 2 (!) monomers of partner 1. This results in partner 2
occupying throughout the cell two different, but mutually exclusive
positions. You could look at it as if the AU has 2 copies of partner 2 each
at approx 50% occupancy and bo!
th overlapping, making the density not very interpretable to say the least.
By carefull analysis of anomalous data we could actually confirm the
position of a heavy atom native to the partner 2 structure. Using reliable
models available for partner 2 and the position of the heavy atom it turns
out we seem to be right: applying either P4 or I4 symmetry makes all partner
2 models superimpose. I think this explains perhaps the very weak P4
reflections, as on a very microscopic scale the crystal is obviously
heterogenous, having partner 2 in either position A or B for all unit cells
resulting in the odd and diffuse reflections seen in between the strong and
sharp I4122 reflections. Overall, the crystal diffracts like it is I4122.
Explaining this particular case has always been difficult without graphics
to illustrate it, but i hope this makes sense.
Needless to say we have returned to find other crystallisation conditions. I
would be very interested to find out if anyone else has ever encountered
similar case(s), and even more if they have found a solution to this kind of
problem.
yours,
David Leys
Crystallography Group,
University Of Leicester
www.le.ac.uk/biochem/pcem1/Xraywebsite/leys.htm
Original Message-----
From: sameeta bilgrami [mailto:sameeta_b@yahoo.co.in]
Sent: Fri 28/03/2003 15:33
To: ccp4bb@dl.ac.uk
Cc:
Subject: [ccp4bb]: spacegroup problem
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
dear all,
hello,
I have asked this question once before, but i
think now i have more data and information and
threfore can present my problem in a better way.
I have a set of datasets collected at
synchroton.( native till 1.9 a resolution and four
derivatives till 2.2 A resolution, Hgcl2 HgCN, Pt and
Pb). There is an ambiguity in spacegroup for these
crystals. Denzo gives P tetragonal if allowed to take
the default no. of spots. It gives I tetragonal if the
no. of spots taken for autoindexing are reduced. Same
is the case with automar. Somebody suggested that i
should take two images 90 degrees apart and see. Two
images taken 90 degree apart does reduce the error ,
but the trick remains the same less no. of spots give
I tetragonal and more no. of spots give P tetragonal.
I then tried mosflm. Here i do not know how to change
the no. of spots to be taken and so I treid with
default no. of spots. It gave varying results. It gave
P tetragonal for the native and the Pt. crytal but
gave I tetragonal for the rest of the derivatives.
I have also calculated native patterson till 2.2
a resolution with the P tetragonal mtz file for all
datasets. All the datasets give very large off origin
peak at 1/2, 1/2, 1/2. The origin peak is on an
average ~110 and the one at 1/2, 1/2, 1/2 is about 88.
I looked at h+k+l=2n+1 reflections in the sca
file of P tetragonal. Many reflections are weak but
many are present.
I have an earlier set of datasets( 1 native
till 2.5 and 3 derivatives also till 2.5) collected in
house. These crystals unambiguosly give the spacegroup
as I tetragonal. The cell parameters of both the
datasets (collected in house and those at synchroton)
are the same.
From the data collected in house i could
manage to build a partial model ( a little less than
half of the structure). The heavy atom derivatives
were not very good and that is why i had tried to
collect another datset at the synchroton. My protein
does not have much regular secondary structure and the
phases also not being good i was not sure wether i had
traced the right path or not. So i used this model for
MR with the native data collected in house and i got a
good enough solution. Emboldened i tried my hand at
the native collected at synchroton. This did not give
the solution in P43212/p41212(as suggested by
sytamatic absences) Neither did it give solution in
any of the other P tetragonal spacegroups. I then
tried the mtz file of I4122 and I did get a solution.
What does this suggest? does this mean that my
real space group is I tetragonal
An off origin peak at 1/2 1/2 1/2 may suggest
pseudotranslational symmetry....but ..what if i
process an I tetrgonal as Ptetragonal wont that also
give a peak at 1/2, 1/2, 1/2. How do we distinguish
between these two cases.....
there are certainly certain changes taking place at
the synchroton that make the data collected there to
have an ambiguity with regards to the spacegroups
though the same crystals give umambiguous spacegroup
with data collected in house. What implication these
changes have with regards to the isomorphism of a set
of datasets collected at synchroton. This is because
the same sort of "change" may not occur for different
crystals. I am sayiing this because though i am
observing large differences between the native and
derivatives, i am ending up with bad phasing
statistics with solve and mlphare ( i have tried
p43212, p41212 and i4122)
What else can i do to build up more of my model
(apart from trying to collect another better set of
derivatives)
Thankyou all for your kind attention. Thanks
in advance for all the suggestions.
Yours sincerely,
Sameeta Bilgrami
________________________________________________________________________
Missed your favourite TV serial last night? Try the new, Yahoo! TV.
visit http://in.tv.yahoo.com
-----Original Message-----
From: sameeta bilgrami [mailto:sameeta_b@yahoo.co.in]
Sent: Fri 28/03/2003 15:33
To: ccp4bb@dl.ac.uk
Cc:
Subject: [ccp4bb]: spacegroup problem
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
dear all,
hello,
I have asked this question once before, but i
think now i have more data and information and
threfore can present my problem in a better way.
I have a set of datasets collected at
synchroton.( native till 1.9 a resolution and four
derivatives till 2.2 A resolution, Hgcl2 HgCN, Pt and
Pb). There is an ambiguity in spacegroup for these
crystals. Denzo gives P tetragonal if allowed to take
the default no. of spots. It gives I tetragonal if the
no. of spots taken for autoindexing are reduced. Same
is the case with automar. Somebody suggested that i
should take two images 90 degrees apart and see. Two
images taken 90 degree apart does reduce the error ,
but the trick remains the same less no. of spots give
I tetragonal and more no. of spots give P tetragonal.
I then tried mosflm. Here i do not know how to change
the no. of spots to be taken and so I treid with
default no. of spots. It gave varying results. It gave
P tetragonal for the native and the Pt. crytal but
gave I tetragonal for the rest of the derivatives.
I have also calculated native patterson till 2.2
a resolution with the P tetragonal mtz file for all
datasets. All the datasets give very large off origin
peak at 1/2, 1/2, 1/2. The origin peak is on an
average ~110 and the one at 1/2, 1/2, 1/2 is about 88.
I looked at h+k+l=2n+1 reflections in the sca
file of P tetragonal. Many reflections are weak but
many are present.
I have an earlier set of datasets( 1 native
till 2.5 and 3 derivatives also till 2.5) collected in
house. These crystals unambiguosly give the spacegroup
as I tetragonal. The cell parameters of both the
datasets (collected in house and those at synchroton)
are the same.
From the data collected in house i could
manage to build a partial model ( a little less than
half of the structure). The heavy atom derivatives
were not very good and that is why i had tried to
collect another datset at the synchroton. My protein
does not have much regular secondary structure and the
phases also not being good i was not sure wether i had
traced the right path or not. So i used this model for
MR with the native data collected in house and i got a
good enough solution. Emboldened i tried my hand at
the native collected at synchroton. This did not give
the solution in P43212/p41212(as suggested by
sytamatic absences) Neither did it give solution in
any of the other P tetragonal spacegroups. I then
tried the mtz file of I4122 and I did get a solution.
What does this suggest? does this mean that my
real space group is I tetragonal
An off origin peak at 1/2 1/2 1/2 may suggest
pseudotranslational symmetry....but ..what if i
process an I tetrgonal as Ptetragonal wont that also
give a peak at 1/2, 1/2, 1/2. How do we distinguish
between these two cases.....
there are certainly certain changes taking place at
the synchroton that make the data collected there to
have an ambiguity with regards to the spacegroups
though the same crystals give umambiguous spacegroup
with data collected in house. What implication these
changes have with regards to the isomorphism of a set
of datasets collected at synchroton. This is because
the same sort of "change" may not occur for different
crystals. I am sayiing this because though i am
observing large differences between the native and
derivatives, i am ending up with bad phasing
statistics with solve and mlphare ( i have tried
p43212, p41212 and i4122)
What else can i do to build up more of my model
(apart from trying to collect another better set of
derivatives)
Thankyou all for your kind attention. Thanks
in advance for all the suggestions.
Yours sincerely,
Sameeta Bilgrami
________________________________________________________________________
Missed your favourite TV serial last night? Try the new, Yahoo! TV.
visit http://in.tv.yahoo.com
LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.