Re: [ccp4bb] distinguish ligand binding sites within a protein

[Bärbel Blaum]

Hi Ed,

you are right about the original question, but what I mean is this: if  
the occupancies (and B-factors) differ so much in crystals with  
IDENTICAL binding sites, i.e. identical affinities, does this not show  
that occupancies (and B-factors) do not reflect affinities alone, but  
equally local packing? There might be individual cases in which such  
effects can be neglected, but generally I think trying to determine  
affinities from crystal soaks is, hmm, not very good pratice, simply  
because there are other dedicated methods to do it that suffer less  
from side effects. Including the docking approach.


Kind regards, Baerbel


Quoting Ed Pozharski :


If I understand the original post correctly, the binding sites in
question are not chemically identical.  While it's possible that lattice
may invert the order in which sites are occupied, it is not very likely
given that affinity gap is sufficient to be observable by ITC.

Mutagenesis is a good option too.

On Tue, 2013-11-19 at 17:12 +0100, Bärbel Blaum wrote:

Hello,

we work with proteins that have typically several chemically identical
binding sites (viral capsid proteins fully assembled or as multimeric
assembly-intermediates). Depending on how long at which concentrations
they are soaked the chemically identical ligand pockets within one
asymmetric unit are typically occupied to different levels purely
because of individual crystal contacts and accessibility. I therefore
think that neither soaking with different concentrations nor B-factor
analysis are solid methods to determine some sort of relative
affinities. I'd suggest to design mutants for either binding site and
ITC measurements with the mutant proteins. This might also tell you if
some sort of co-op exists between both sites.

Baerbel

Quoting Ed Pozharski :

> IMHO, while explaining binding affinity from a structure is fun, it does
> not prove anything.  Assuming that I understand your situation
> correctly, you can (relatively) easily find out from experiment which
> pocket has higher affinity.  Just do soaks with different ligand
> concentrations - the expectation is that the weaker binding site will
> become partially occupied first.
>
> On Tue, 2013-11-19 at 04:58 +, Xiaodi Yu wrote:
>> Hi Wei:
>>
>> Based on the structure, you can calculate the binding surface between
>> the protein and the ligand. Maybe the two binding pockets will give
>> you two different numbers. And the larger one usually can have the
>> higher binding affinity.  You also can analyse how the ligand
>> interacts with the protein though hydrophobic or electrostatic
>> interaction , etc?  the last, you may also compare the b factors of
>> the ligand or the protein binding pocket regions after you refining
>> the structure. These things may give you some hints about which
>> binding site is more strong.
>>
>> Dee
>>
>>
>> __
>> Date: Mon, 18 Nov 2013 22:45:58 -0500
>> From: wei.shi...@gmail.com
>> Subject: Re: [ccp4bb] distinguish ligand binding sites within a
>> protein
>> To: CCP4BB@JISCMAIL.AC.UK
>>
>> Thank you so much for the suggestions, Tomas! Yes, my ligand is a
>> small molecule. I have the crystal structure of the ligands bound to
>> the protein, do I still need to computationally dock the ligand to the
>> two pockets, can I calculate the parameters of binding directly using
>> the crystal structure?
>>
>> Best,
>> Wei
>>
>>
>>
>> On Mon, Nov 18, 2013 at 9:03 PM, Tomas Malinauskas
>>  wrote:
>> Dear Wei Shi,
>> is your ligand a small molecule? If it is a small molecule, I
>> would
>> try to computationally dock the small molecule to two pockets
>> separately using AutoDock, and look at the estimated free
>> energies of
>> binding.
>> Best wishes,
>> Tomas
>>
>> On Mon, Nov 18, 2013 at 8:55 PM, Wei Shi
>>  wrote:
>> > Hi all,
>> > I got the crystal structure of a transcription factor, and
>> every monomer
>> > binds two molecules of the same ligand in different binding
>> pockets. And I
>> > also did the ITC experiment, titrating the ligand into the
>> protein, and got
>> > a U-shaped curve. The binding affinity for the first binding
>> site is higher
>> > than the second binding site.
>> > I am wondering whether I could computationally determine
>> from the
>> > protein-ligand complex structure that which binding site has
>> higher affinity
>> > for the ligand and correlate the binding sites with the
>> parameters I got
>> > from ITC experiment.
>> > Thank you so much!
>> >
>> > Best,
>> > Wei
>>
>>
>>
>
> --
> Edwin Pozharski, PhD, Assistant Professor
> University of Maryland, Baltimore
> --
> When the Way is forgotten duty and 

Re: [ccp4bb] distinguish ligand binding sites within a protein

[Ed Pozharski]

If I understand the original post correctly, the binding sites in
question are not chemically identical.  While it's possible that lattice
may invert the order in which sites are occupied, it is not very likely
given that affinity gap is sufficient to be observable by ITC.

Mutagenesis is a good option too.

On Tue, 2013-11-19 at 17:12 +0100, Bärbel Blaum wrote:
> Hello,
> 
> we work with proteins that have typically several chemically identical  
> binding sites (viral capsid proteins fully assembled or as multimeric  
> assembly-intermediates). Depending on how long at which concentrations  
> they are soaked the chemically identical ligand pockets within one  
> asymmetric unit are typically occupied to different levels purely  
> because of individual crystal contacts and accessibility. I therefore  
> think that neither soaking with different concentrations nor B-factor  
> analysis are solid methods to determine some sort of relative  
> affinities. I'd suggest to design mutants for either binding site and  
> ITC measurements with the mutant proteins. This might also tell you if  
> some sort of co-op exists between both sites.
> 
> Baerbel
> 
> Quoting Ed Pozharski :
> 
> > IMHO, while explaining binding affinity from a structure is fun, it does
> > not prove anything.  Assuming that I understand your situation
> > correctly, you can (relatively) easily find out from experiment which
> > pocket has higher affinity.  Just do soaks with different ligand
> > concentrations - the expectation is that the weaker binding site will
> > become partially occupied first.
> >
> > On Tue, 2013-11-19 at 04:58 +, Xiaodi Yu wrote:
> >> Hi Wei:
> >>
> >> Based on the structure, you can calculate the binding surface between
> >> the protein and the ligand. Maybe the two binding pockets will give
> >> you two different numbers. And the larger one usually can have the
> >> higher binding affinity.  You also can analyse how the ligand
> >> interacts with the protein though hydrophobic or electrostatic
> >> interaction , etc?  the last, you may also compare the b factors of
> >> the ligand or the protein binding pocket regions after you refining
> >> the structure. These things may give you some hints about which
> >> binding site is more strong.
> >>
> >> Dee
> >>
> >>
> >> __
> >> Date: Mon, 18 Nov 2013 22:45:58 -0500
> >> From: wei.shi...@gmail.com
> >> Subject: Re: [ccp4bb] distinguish ligand binding sites within a
> >> protein
> >> To: CCP4BB@JISCMAIL.AC.UK
> >>
> >> Thank you so much for the suggestions, Tomas! Yes, my ligand is a
> >> small molecule. I have the crystal structure of the ligands bound to
> >> the protein, do I still need to computationally dock the ligand to the
> >> two pockets, can I calculate the parameters of binding directly using
> >> the crystal structure?
> >>
> >> Best,
> >> Wei
> >>
> >>
> >>
> >> On Mon, Nov 18, 2013 at 9:03 PM, Tomas Malinauskas
> >>  wrote:
> >> Dear Wei Shi,
> >> is your ligand a small molecule? If it is a small molecule, I
> >> would
> >> try to computationally dock the small molecule to two pockets
> >> separately using AutoDock, and look at the estimated free
> >> energies of
> >> binding.
> >> Best wishes,
> >> Tomas
> >>
> >> On Mon, Nov 18, 2013 at 8:55 PM, Wei Shi
> >>  wrote:
> >> > Hi all,
> >> > I got the crystal structure of a transcription factor, and
> >> every monomer
> >> > binds two molecules of the same ligand in different binding
> >> pockets. And I
> >> > also did the ITC experiment, titrating the ligand into the
> >> protein, and got
> >> > a U-shaped curve. The binding affinity for the first binding
> >> site is higher
> >> > than the second binding site.
> >> > I am wondering whether I could computationally determine
> >> from the
> >> > protein-ligand complex structure that which binding site has
> >> higher affinity
> >> > for the ligand and correlate the binding sites with the
> >> parameters I got
> >> > from ITC experiment.
> >> > Thank you so much!
> >> >
> >> > Best,
> >> > Wei
> >>
> >>
> >>
> >
> > --
> > Edwin Pozharski, PhD, Assistant Professor
> > University of Maryland, Baltimore
> > --
> > When the Way is forgotten duty and justice appear;
> > Then knowledge and wisdom are born along with hypocrisy.
> > When harmonious relationships dissolve then respect and devotion arise;
> > When a nation falls to chaos then loyalty and patriotism are born.
> > --   / Lao Tse /
> >
> 
> 
> 

-- 
Edwin Pozharski, PhD, Assistant Professor
University of Maryland, Baltimore
--
When the Way is forgotten duty and jus

[ccp4bb] PostDoc - Structural Biology - Cancer

[Alessandro Vannini]

Dear CCP4ers,

a post-doctoral position is available immediately in the laboratory of Dr. 
Alessandro Vannini, within the Division of Structural Biology at The Institute 
of Cancer Research in Chelsea, London, UK. We are looking for highly motivated 
individuals with a strong interest in structural characterization of 
multi-subunit macromolecular complexes. We currently employ single particle 
electron microscopy analysis and X-ray crystallography, as well as biochemical 
and biophysical analysis, to elucidate the role of the RNA Polymerase III and 
associated factors in cancer development.

The Division of Structural Biology has managed facilities for protein 
crystallography (Bruker Microstar and CCD detector and crystallisation robots), 
cryo-electron microscopy (FEI Tecnai F20 and T12), and protein production with 
expertise in multi-subunit expression (insect cell, yeast and bacterial 
expression, including a 60 L fermentor). The Division is also well equipped 
with equipment for biophysical analysis (e.g. ITC, fluorescence, multi-angle 
light scattering).

Applicants should possess a PhD (or equivalent) in biochemistry or molecular 
biology and a sound knowledge in production and purification of macromolecular 
complexes for structural biology analysis. Previous experience with 
purification of multi-subunit complexes and/or biochemistry of protein-nucleic 
acid complexes would be beneficial. Previous experience in protein 
crystallography and/or single particle electron microscopy would be desirable. 

Appointment will be on Fixed Term Contract for 3 years in the first instance, 
with a starting salary in the range of £28,425 to £34,944 p.a. inclusive (based 
on previous post-doctoral experience).  

CV and covering letter must be submitted online at 
http://www.icr.ac.uk/jobsearch

Job Ref. No. 1378320
Closing Date: 29th November 2014

For informal inquiries you can contact me directly at 
alessandro.vann...@icr.ac.uk

Re: [ccp4bb] distinguish ligand binding sites within a protein

[Bärbel Blaum]

Hello,

we work with proteins that have typically several chemically identical  
binding sites (viral capsid proteins fully assembled or as multimeric  
assembly-intermediates). Depending on how long at which concentrations  
they are soaked the chemically identical ligand pockets within one  
asymmetric unit are typically occupied to different levels purely  
because of individual crystal contacts and accessibility. I therefore  
think that neither soaking with different concentrations nor B-factor  
analysis are solid methods to determine some sort of relative  
affinities. I'd suggest to design mutants for either binding site and  
ITC measurements with the mutant proteins. This might also tell you if  
some sort of co-op exists between both sites.


Baerbel

Quoting Ed Pozharski :


IMHO, while explaining binding affinity from a structure is fun, it does
not prove anything.  Assuming that I understand your situation
correctly, you can (relatively) easily find out from experiment which
pocket has higher affinity.  Just do soaks with different ligand
concentrations - the expectation is that the weaker binding site will
become partially occupied first.

On Tue, 2013-11-19 at 04:58 +, Xiaodi Yu wrote:

Hi Wei:

Based on the structure, you can calculate the binding surface between
the protein and the ligand. Maybe the two binding pockets will give
you two different numbers. And the larger one usually can have the
higher binding affinity.  You also can analyse how the ligand
interacts with the protein though hydrophobic or electrostatic
interaction , etc?  the last, you may also compare the b factors of
the ligand or the protein binding pocket regions after you refining
the structure. These things may give you some hints about which
binding site is more strong.

Dee


__
Date: Mon, 18 Nov 2013 22:45:58 -0500
From: wei.shi...@gmail.com
Subject: Re: [ccp4bb] distinguish ligand binding sites within a
protein
To: CCP4BB@JISCMAIL.AC.UK

Thank you so much for the suggestions, Tomas! Yes, my ligand is a
small molecule. I have the crystal structure of the ligands bound to
the protein, do I still need to computationally dock the ligand to the
two pockets, can I calculate the parameters of binding directly using
the crystal structure?

Best,
Wei



On Mon, Nov 18, 2013 at 9:03 PM, Tomas Malinauskas
 wrote:
Dear Wei Shi,
is your ligand a small molecule? If it is a small molecule, I
would
try to computationally dock the small molecule to two pockets
separately using AutoDock, and look at the estimated free
energies of
binding.
Best wishes,
Tomas

On Mon, Nov 18, 2013 at 8:55 PM, Wei Shi
 wrote:
> Hi all,
> I got the crystal structure of a transcription factor, and
every monomer
> binds two molecules of the same ligand in different binding
pockets. And I
> also did the ITC experiment, titrating the ligand into the
protein, and got
> a U-shaped curve. The binding affinity for the first binding
site is higher
> than the second binding site.
> I am wondering whether I could computationally determine
from the
> protein-ligand complex structure that which binding site has
higher affinity
> for the ligand and correlate the binding sites with the
parameters I got
> from ITC experiment.
> Thank you so much!
>
> Best,
> Wei





--
Edwin Pozharski, PhD, Assistant Professor
University of Maryland, Baltimore
--
When the Way is forgotten duty and justice appear;
Then knowledge and wisdom are born along with hypocrisy.
When harmonious relationships dissolve then respect and devotion arise;
When a nation falls to chaos then loyalty and patriotism are born.
--   / Lao Tse /





--
Bärbel Blaum, Ph.D.
Interfakultäres Institut für Biochemie (IFIB)
Hoppe-Seyler-Strasse 4
D-72076 Tübingen
Germany
+49 70 71 29 73 375

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Kay Diederichs]

Dear Ed,

when it comes to deciding about the high-resolution cutoff, I agree that the 
paired-refinement technique should be used - even more so as it does not 
require any of the data quality indicators!

But my posting was meant in a more general sense (i.e. not only talking about 
high-resol cutoff): crystallography is complicated enough that we need the 
right means and tools, which should enable us to make the right decisions. For 
instance, we need to decide which  detector distance to choose, which 
oscillation range to use, at which frame to cut in case of radiation damage, 
what to do in case of anisotropy, which data to merge, which datasets to choose 
for phasing, MR, refinement ... ... .

The tools, means and decisions need to be based on the right concepts. I find 
that the historical concepts (centered on Rmerge), when it comes to data 
quality indicators, are quite awkward and do not enable understanding of the 
problems, rather, they obscure the solutions. I hope that you find the concepts 
that I mention logical and helpful.

best,

Kay


On Tue, 19 Nov 2013 10:27:16 -0500, Ed Pozharski  wrote:

>Dear Kay,
>
>I wonder what is your opinion of the following proposition.
>
>"None of the data quality indicators derived from data alone matter too
>much".
>
>Let me explain what I mean by this.
>
>Ultimately, I truly don't care what value of Rmerge, Rpim, or even CC1/2
>data processing produces from the set of frames I toss at it.  Surely it
>is important to keep an eye on them to verify that dataset is kosher and
>to obtain an *initial* estimate of the resolution limit of useful data.
>But the actual values never deter from trying to solve the structure.
>If it cannot be solved - well, then none of the aforementioned
>indicators matter at all.  If it is solved, the only remaining question
>is how far the useful data goes.  And that should be determined using
>Karplus-Diederichs (KD) test.  I do use CC1/2~0.5 as initial resolution
>cutoff these days, but before finalizing any model I run the KD test.  I
>do look at where the Rpim, I/sigma and CC1/2 end up at the resolution
>edge, but only out of curiosity and to adjust my perception of how they
>correlate with true resolution.
>
>And I think that efforts should be targeted at optimizing KD test as a
>tool rather than being distracted by outdated approaches that were
>proposed in the computationally-handicapped times.
>
>It is entirely possible that all this is exactly what you said, just
>with different wording.  But I guess more wording is still needed given
>that people keep asking about Rmerge.
>
>Cheers,
>
>Ed.
>
>On Tue, 2013-11-19 at 14:22 +, Kay Diederichs wrote:
>> Hi Jim,
>>
>> of course the issue of crystallographic data quality indicators deserves a 
>> somewhat more appropriate (or at least more permanent, and peer-reviewed) 
>> means of dissemination than CCP4BB. Nevertheless I'll sum up some of the 
>> most important points I can think of:
>>
>> A) all data quality indicators measure precision, not accuracy
>> B) there are those data quality indicators that measure the precision of 
>> unmerged data:
>> (Rsym=)Rmerge, Rmeas, (I/sigma)_unmerged
>> and those that measure the precision of merged data:
>> Rpim, Rsplit (the FEL community uses this; same as R_mrgd_I - see 
>> Diederichs&Karplus 1997), CC1/2, (I/sigma)_merged
>> The merged indicators usually differ by a factor of sqrt(m) from their 
>> unmerged counterparts, where m is multiplicity. Rsplit (~R_mrgd_I) and CC1/2 
>> compare random half-datasets which may be more robust than just hoping that 
>> the explicit sqrt(m) law holds (it only holds for unrelated errors). There 
>> is no unmerged counterpart of CC1/2.
>> C) Since downstream steps use intensities, it is preferable to use a data 
>> quality indicator that does not require sigma to be estimated, because the 
>> authors of the different data processing programs/algorithms have different 
>> ideas how this should be done. This rules out I/sigma as a useful quality 
>> indicator - at least as soon as different programs look at the same data.
>> D) Merged data quality indicators are more useful because we are using 
>> merged data for downstream steps (phasing, molecular replacement, 
>> refinement), so we need to know _their_ precision, not that of the unmerged 
>> data.
>> E) Rpim and Rsplit are calculated from intensities and have a different 
>> asymptotic behaviour than model R-values (Rwork, Rfree), so they cannot be 
>> meaningfully be compared with model R-values (i.e. their numerical value 
>> tells you nothing about the Rwork/Rfree your model can be refined to). This 
>> is very different from CC1/2 - it can be used to calculate CC*, a quantity 
>> that is the upper limit of what the CC of the model intensities against the 
>> experimental intensities can reach.
>>
>> I'll stop here. Most of this may be at variance with what we were all 
>> brought up with, but it's time for a change!
>>
>> best,
>>
>> Kay
>>

Re: [ccp4bb] distinguish ligand binding sites within a protein

[Ed Pozharski]

IMHO, while explaining binding affinity from a structure is fun, it does
not prove anything.  Assuming that I understand your situation
correctly, you can (relatively) easily find out from experiment which
pocket has higher affinity.  Just do soaks with different ligand
concentrations - the expectation is that the weaker binding site will
become partially occupied first.

On Tue, 2013-11-19 at 04:58 +, Xiaodi Yu wrote:
> Hi Wei:
> 
> Based on the structure, you can calculate the binding surface between
> the protein and the ligand. Maybe the two binding pockets will give
> you two different numbers. And the larger one usually can have the
> higher binding affinity.  You also can analyse how the ligand
> interacts with the protein though hydrophobic or electrostatic
> interaction , etc?  the last, you may also compare the b factors of
> the ligand or the protein binding pocket regions after you refining
> the structure. These things may give you some hints about which
> binding site is more strong.
> 
> Dee
> 
> 
> __
> Date: Mon, 18 Nov 2013 22:45:58 -0500
> From: wei.shi...@gmail.com
> Subject: Re: [ccp4bb] distinguish ligand binding sites within a
> protein
> To: CCP4BB@JISCMAIL.AC.UK
> 
> Thank you so much for the suggestions, Tomas! Yes, my ligand is a
> small molecule. I have the crystal structure of the ligands bound to
> the protein, do I still need to computationally dock the ligand to the
> two pockets, can I calculate the parameters of binding directly using
> the crystal structure? 
> 
> Best,
> Wei 
> 
> 
> 
> On Mon, Nov 18, 2013 at 9:03 PM, Tomas Malinauskas
>  wrote:
> Dear Wei Shi,
> is your ligand a small molecule? If it is a small molecule, I
> would
> try to computationally dock the small molecule to two pockets
> separately using AutoDock, and look at the estimated free
> energies of
> binding.
> Best wishes,
> Tomas
> 
> On Mon, Nov 18, 2013 at 8:55 PM, Wei Shi
>  wrote:
> > Hi all,
> > I got the crystal structure of a transcription factor, and
> every monomer
> > binds two molecules of the same ligand in different binding
> pockets. And I
> > also did the ITC experiment, titrating the ligand into the
> protein, and got
> > a U-shaped curve. The binding affinity for the first binding
> site is higher
> > than the second binding site.
> > I am wondering whether I could computationally determine
> from the
> > protein-ligand complex structure that which binding site has
> higher affinity
> > for the ligand and correlate the binding sites with the
> parameters I got
> > from ITC experiment.
> > Thank you so much!
> >
> > Best,
> > Wei
> 
> 
> 

-- 
Edwin Pozharski, PhD, Assistant Professor
University of Maryland, Baltimore
--
When the Way is forgotten duty and justice appear;
Then knowledge and wisdom are born along with hypocrisy.
When harmonious relationships dissolve then respect and devotion arise;
When a nation falls to chaos then loyalty and patriotism are born.
--   / Lao Tse /

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Ed Pozharski]

Dear Kay,

I wonder what is your opinion of the following proposition.

"None of the data quality indicators derived from data alone matter too
much".

Let me explain what I mean by this.

Ultimately, I truly don't care what value of Rmerge, Rpim, or even CC1/2
data processing produces from the set of frames I toss at it.  Surely it
is important to keep an eye on them to verify that dataset is kosher and
to obtain an *initial* estimate of the resolution limit of useful data.
But the actual values never deter from trying to solve the structure.
If it cannot be solved - well, then none of the aforementioned
indicators matter at all.  If it is solved, the only remaining question
is how far the useful data goes.  And that should be determined using
Karplus-Diederichs (KD) test.  I do use CC1/2~0.5 as initial resolution
cutoff these days, but before finalizing any model I run the KD test.  I
do look at where the Rpim, I/sigma and CC1/2 end up at the resolution
edge, but only out of curiosity and to adjust my perception of how they
correlate with true resolution.

And I think that efforts should be targeted at optimizing KD test as a
tool rather than being distracted by outdated approaches that were
proposed in the computationally-handicapped times.

It is entirely possible that all this is exactly what you said, just
with different wording.  But I guess more wording is still needed given
that people keep asking about Rmerge.  

Cheers,

Ed.

On Tue, 2013-11-19 at 14:22 +, Kay Diederichs wrote:
> Hi Jim,
> 
> of course the issue of crystallographic data quality indicators deserves a 
> somewhat more appropriate (or at least more permanent, and peer-reviewed) 
> means of dissemination than CCP4BB. Nevertheless I'll sum up some of the most 
> important points I can think of:
> 
> A) all data quality indicators measure precision, not accuracy
> B) there are those data quality indicators that measure the precision of 
> unmerged data:
> (Rsym=)Rmerge, Rmeas, (I/sigma)_unmerged 
> and those that measure the precision of merged data:
> Rpim, Rsplit (the FEL community uses this; same as R_mrgd_I - see 
> Diederichs&Karplus 1997), CC1/2, (I/sigma)_merged
> The merged indicators usually differ by a factor of sqrt(m) from their 
> unmerged counterparts, where m is multiplicity. Rsplit (~R_mrgd_I) and CC1/2 
> compare random half-datasets which may be more robust than just hoping that 
> the explicit sqrt(m) law holds (it only holds for unrelated errors). There is 
> no unmerged counterpart of CC1/2.
> C) Since downstream steps use intensities, it is preferable to use a data 
> quality indicator that does not require sigma to be estimated, because the 
> authors of the different data processing programs/algorithms have different 
> ideas how this should be done. This rules out I/sigma as a useful quality 
> indicator - at least as soon as different programs look at the same data. 
> D) Merged data quality indicators are more useful because we are using merged 
> data for downstream steps (phasing, molecular replacement, refinement), so we 
> need to know _their_ precision, not that of the unmerged data.
> E) Rpim and Rsplit are calculated from intensities and have a different 
> asymptotic behaviour than model R-values (Rwork, Rfree), so they cannot be 
> meaningfully be compared with model R-values (i.e. their numerical value 
> tells you nothing about the Rwork/Rfree your model can be refined to). This 
> is very different from CC1/2 - it can be used to calculate CC*, a quantity 
> that is the upper limit of what the CC of the model intensities against the 
> experimental intensities can reach. 
> 
> I'll stop here. Most of this may be at variance with what we were all brought 
> up with, but it's time for a change!
> 
> best,
> 
> Kay
> 
> On Tue, 19 Nov 2013 13:18:19 +, Jim Pflugrath  
> wrote:
> 
> >Graeme wrote:
> >"... Rpim is much more instructive. ... as each of these tells something 
> >different."
> >
> >I have to ask:
> >"Why is Rpim much more instructive?  I'm trying to figure this out still.  
> >Can one please summarize what are best practices with all these numbers and 
> >how each of these tells something different?"
> >
> >Another problem that I see is that folks can adjust their sigmas many 
> >different ways without knowing they have adjusted their sigmas.  And they 
> >can be adjusted incorrectly when they are adjusted.
> >
> >BTW, Graeme is correct about lots of multiple low I/sigI observations for 
> >each Bragg reflection in a resolution shell will lead to 100% (or higher) 
> >Rmerge with  of 3.  This assumes no systematic errors and only 
> >randomly distributed random errors (a rare if not impossible situation, I 
> >would think).  I will defer to others about what the relevance of that is.
> >
> >Thanks for any insights, Jim
> >
> >
> >
> >From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Graeme Winter 
> >[graeme.win...@gmail.com]
> >Sent: Tue

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Kay Diederichs]

Hi Jim,

of course the issue of crystallographic data quality indicators deserves a 
somewhat more appropriate (or at least more permanent, and peer-reviewed) means 
of dissemination than CCP4BB. Nevertheless I'll sum up some of the most 
important points I can think of:

A) all data quality indicators measure precision, not accuracy
B) there are those data quality indicators that measure the precision of 
unmerged data:
(Rsym=)Rmerge, Rmeas, (I/sigma)_unmerged 
and those that measure the precision of merged data:
Rpim, Rsplit (the FEL community uses this; same as R_mrgd_I - see 
Diederichs&Karplus 1997), CC1/2, (I/sigma)_merged
The merged indicators usually differ by a factor of sqrt(m) from their unmerged 
counterparts, where m is multiplicity. Rsplit (~R_mrgd_I) and CC1/2 compare 
random half-datasets which may be more robust than just hoping that the 
explicit sqrt(m) law holds (it only holds for unrelated errors). There is no 
unmerged counterpart of CC1/2.
C) Since downstream steps use intensities, it is preferable to use a data 
quality indicator that does not require sigma to be estimated, because the 
authors of the different data processing programs/algorithms have different 
ideas how this should be done. This rules out I/sigma as a useful quality 
indicator - at least as soon as different programs look at the same data. 
D) Merged data quality indicators are more useful because we are using merged 
data for downstream steps (phasing, molecular replacement, refinement), so we 
need to know _their_ precision, not that of the unmerged data.
E) Rpim and Rsplit are calculated from intensities and have a different 
asymptotic behaviour than model R-values (Rwork, Rfree), so they cannot be 
meaningfully be compared with model R-values (i.e. their numerical value tells 
you nothing about the Rwork/Rfree your model can be refined to). This is very 
different from CC1/2 - it can be used to calculate CC*, a quantity that is the 
upper limit of what the CC of the model intensities against the experimental 
intensities can reach. 

I'll stop here. Most of this may be at variance with what we were all brought 
up with, but it's time for a change!

best,

Kay

On Tue, 19 Nov 2013 13:18:19 +, Jim Pflugrath  
wrote:

>Graeme wrote:
>"... Rpim is much more instructive. ... as each of these tells something 
>different."
>
>I have to ask:
>"Why is Rpim much more instructive?  I'm trying to figure this out still.  Can 
>one please summarize what are best practices with all these numbers and how 
>each of these tells something different?"
>
>Another problem that I see is that folks can adjust their sigmas many 
>different ways without knowing they have adjusted their sigmas.  And they can 
>be adjusted incorrectly when they are adjusted.
>
>BTW, Graeme is correct about lots of multiple low I/sigI observations for each 
>Bragg reflection in a resolution shell will lead to 100% (or higher) Rmerge 
>with  of 3.  This assumes no systematic errors and only randomly 
>distributed random errors (a rare if not impossible situation, I would think). 
> I will defer to others about what the relevance of that is.
>
>Thanks for any insights, Jim
>
>
>
>From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Graeme Winter 
>[graeme.win...@gmail.com]
>Sent: Tuesday, November 19, 2013 2:02 AM
>To: CCP4BB@JISCMAIL.AC.UK
>Subject: Re: [ccp4bb] 100% Rmerge in high resolution shell
>
>Usually this means that you have relatively high multiplicity, which 
>give-or-take improves the I/sig(I) by sqrt(m) where m is the multiplicity, but 
>also increases the Rmerge.
>
>For any given narrow shell of reflections,
>
>Rmerge ~ 0.8 / unmerged(I/sig(I))
>
>merged(I/sig(I)) ~ sqrt(m) * unmerged(I/sig(I))
>
>So it is perfectly possible to have unmerged I/sig(I) of 0.8 which will give 
>you an Rmerge of around 1.0, and have I/sig(I) (merged) around 3, by having 
>multiplciity 14 or so. I suggest that this is the case: if it is much lower 
>than this there is something odd going on.
>
>For the merged I/sig(I) Rpim is much more instructive. I'd love it if people 
>reported merged and unmerged I/sig(I), Rmerge, Rmeas, Rpim, CC1/2, ... as each 
>of these tells something different.
>
>Best wishes,
>
>Graeme
>
>Possibly useful papers:
>
>http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html
>http://scripts.iucr.org/cgi-bin/paper?he0191
>http://scripts.iucr.org/cgi-bin/paper?he0268
>
>
>
>
>On 19 November 2013 06:43, Shanti Pal Gangwar 
>mailto:gangwar...@gmail.com>> wrote:
>Dear  All
>
>
>Can anyone explain the meaning and relevance of data when the Rmerge is 100% 
>in high resolution shell and I/sig(I) is 3.
>
>
>
>Thanks
>
>
>
>--
>
>regards
>Shanti Pal Gangwar
>School of Life Sciences
>Jawaharlal Nehru University
>New Delhi-110067
>India
>Email:gangwar...@gmail.com
>
>
>
>

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Jim Pflugrath]

Graeme wrote:
"... Rpim is much more instructive. ... as each of these tells something 
different."

I have to ask:
"Why is Rpim much more instructive?  I'm trying to figure this out still.  Can 
one please summarize what are best practices with all these numbers and how 
each of these tells something different?"

Another problem that I see is that folks can adjust their sigmas many different 
ways without knowing they have adjusted their sigmas.  And they can be adjusted 
incorrectly when they are adjusted.

BTW, Graeme is correct about lots of multiple low I/sigI observations for each 
Bragg reflection in a resolution shell will lead to 100% (or higher) Rmerge 
with  of 3.  This assumes no systematic errors and only randomly 
distributed random errors (a rare if not impossible situation, I would think).  
I will defer to others about what the relevance of that is.

Thanks for any insights, Jim



From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Graeme Winter 
[graeme.win...@gmail.com]
Sent: Tuesday, November 19, 2013 2:02 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] 100% Rmerge in high resolution shell

Usually this means that you have relatively high multiplicity, which 
give-or-take improves the I/sig(I) by sqrt(m) where m is the multiplicity, but 
also increases the Rmerge.

For any given narrow shell of reflections,

Rmerge ~ 0.8 / unmerged(I/sig(I))

merged(I/sig(I)) ~ sqrt(m) * unmerged(I/sig(I))

So it is perfectly possible to have unmerged I/sig(I) of 0.8 which will give 
you an Rmerge of around 1.0, and have I/sig(I) (merged) around 3, by having 
multiplciity 14 or so. I suggest that this is the case: if it is much lower 
than this there is something odd going on.

For the merged I/sig(I) Rpim is much more instructive. I'd love it if people 
reported merged and unmerged I/sig(I), Rmerge, Rmeas, Rpim, CC1/2, ... as each 
of these tells something different.

Best wishes,

Graeme

Possibly useful papers:

http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html
http://scripts.iucr.org/cgi-bin/paper?he0191
http://scripts.iucr.org/cgi-bin/paper?he0268




On 19 November 2013 06:43, Shanti Pal Gangwar 
mailto:gangwar...@gmail.com>> wrote:
Dear  All


Can anyone explain the meaning and relevance of data when the Rmerge is 100% in 
high resolution shell and I/sig(I) is 3.



Thanks



--

regards
Shanti Pal Gangwar
School of Life Sciences
Jawaharlal Nehru University
New Delhi-110067
India
Email:gangwar...@gmail.com

Re: [ccp4bb] Fix cell dimensions

[Kay Diederichs]

Dear Niu,

concerning XDS: there is no way, using just XDS.INP, to force the IDXREF step 
to use a cell that is not compatible with the reflections that COLSPOT found.
But I can think of two ways to reach your goal (even if I believe that 
crystallographically it makes no sense to ignore every second existing 
reflection).

a) stop after the IDXREF step that uses all reflections, and gives you the 40A 
axis that you want to halve (suppose it is the "a" axis). Next, edit SPOT.XDS 
and remove 
all lines which have (in this example) odd h. This can be accomplished with
mv SPOT.XDS SPOT.XDS.all
awk '{if ($5%2==0) print $0}' SPOT.XDS.all > SPOT.XDS 
Then, run IDXREF and subsequent steps - they will use the short a axis.

b) stop after the IDXREF step that uses all reflections, and gives you the 40A 
axis that you want to halve (suppose it is the "a" axis). Next, edit two lines 
in XPARM.XDS: line 4, which has the cell parameters (halve a in this line), and 
line 5, which has the direction of the a axis in space (halve all 3 values). 
Then, run DEFPIX  and subsequent steps - they will use the short a axis.

Hope this helps,

Kay


On Mon, 18 Nov 2013 17:03:06 -0500, Niu Tou  wrote:

>Dear Andrew,
>
>As previously I posted a MR case which has a significant 95% off origin
>peak, some experts suggested to reprocess the data with cutting one axis to
>half, from 40A to 20A. I tried HKL2000 and XDS, none of them is willing to
>give a solution with 20A, even I specify it in XDS script. So I wonder is
>there any way to force this work to be done. Thanks!
>
>Best,
>Niu
>
>
>On Mon, Nov 18, 2013 at 4:56 PM, Andrew Leslie wrote:
>
>> Dear Niu,
>>
>> It depends on which part of processing you are referring
>> to, i.e. the indexing step or the integration step. In MOSFLM there is no
>> way to enforce cell dimensions during indexing, but providing there is an
>> indexing solution that has cell dimensions close to the ones you want, you
>> can enforce a (slightly) different set of cell dimensions during the
>> integration step. Normally other refined parameters will ensure that you
>> still get a good prediction of spot positions.
>>
>> I suspect that this can be done in other programs too.
>>
>> Without knowing why you want to do this, I cannot comment on whether this
>> is the best procedure to follow.
>>
>> Best wishes,
>>
>> Andrew
>>
>>
>>
>> On 18 Nov 2013, at 21:48, Niu Tou  wrote:
>>
>> > Dear All,
>> >
>> > Does any one know how to strictly fix the cell dimensions during data
>> processing? In HKL2000 there is only a keyword to define the longest
>> vector. In XDS there is a option to input cell parameters, but sometimes
>> the program would not follow the input values
>> > and switch back to the one it thinks best. Any suggestions will be
>> appreciated. Thanks!
>> >
>> > Best,
>> > Niu
>>
>>
>

Re: [ccp4bb] translational pseudo symmetry

[Eleanor Dodson]

You would get a different MR solution in P41212 than in P43212 so you
shouldnt test the SAME pdb in both SGS?
Not sure I am understanding this though.
Eleanor

On 19 November 2013 05:02, #CHEN DAN#  wrote:
> Hi Eleanor,
>
> I checked P43212 and P41212 by changing the header of mtz file and running 
> refmac for the same PDB input. P43212 is a better match than P41212.
>
> Sincerely,
> Dan
>
> 
> From: CCP4 bulletin board  on behalf of Eleanor Dodson 
> 
> Sent: Monday, November 18, 2013 8:47 PM
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] translational pseudo symmetry
>
> I guess you have checked that P43212 is a better match than P41212?
> (And that you are running REFMAC against an mtz file with the same
> symmetry as the input PDB - you may need to change the SG in the mtz
> header by hand.
> mtzutils hklin P41212.mtz hklout P43212.mtz
> symm P43212
> end
>
> Or vice versa..
>
> Sorry - THIS IS CRAZY but there you are..
>
> Re the pseudo translation -Randy summs up the situation very clearly.
> I would build my model by hand actually but I am sure PHASER  does itwell too!
>
>
> Something I dont understand but maybe it is to do with your patterson 
> sampling.
>
>
> Peak 3 is a consequence of Pk 1 and Pk2 -
>  Pk 5 is the consequence of Pk 1 and Pk4
> but the peak heights dont exactly fit..
>
> Eleanor
>
> On 18 November 2013 10:19, Randy Read  wrote:
>> Dear Dan,
>>
>> First, you don't want to reprocess in the smaller cell.  What xtriage is
>> saying is that, if *and only if* the translation detected in the Patterson
>> map were an exact crystallographic translation, then you would get the
>> smaller cell.  However, in order for that to be a plausible hypothesis, the
>> Patterson peaks would have to be near to 100% of the origin peak.
>>
>> You actually seem to have a very interesting case, where the Patterson peaks
>> are related by multiples of approximately the same translation.  If you take
>> a translation of 1/2,1/2,1/6 and multiply it by 1, 2 and 3, you get
>> something close to the three biggest peaks in your Patterson (taking account
>> of lattice translations), and these are related by the Patterson inversion
>> centre to what you get if you multiply by 4 and 5.  So the six molecules
>> should be related to each other by something close to a repeated translation
>> of 1/2,1/2,1/6.  (You should check this in the solution that you already
>> have.)  If this were exact, you would have a smaller cell, but it's not
>> exact, and one way in which it is not exact is that the translations along z
>> are not exactly multiples of 1/6.
>>
>> This is reminiscent of a structure that we recently collaborated with
>> Mariusz Jaskolski and Zbyszek Dauter to solve (paper accepted for
>> publication in Acta D).  In that case, there are seven translations of
>> approximately 0,0,1/7.  The difficulty with cases like this is figuring out
>> how to break the exact symmetry.  Any solution that has approximately the
>> right translations will basically fit the data, but you need to find the
>> right combination of deviations from the exact symmetry to get an optimal
>> answer.  If you get the wrong deviations from exact symmetry, the refinement
>> will stall, and this may be the problem that you're facing.
>>
>> You can deal with problems like this in Phaser by using the TNCS NMOL 6
>> command (to say that there are 6 copies related by repeated applications of
>> the same translation).  You should tell Phaser to use the 1/2,1/2,0.174
>> vector (TNCS TRA VECTOR 0.5 0.5 0.174), and hopefully this will break the
>> symmetry in a way that subsequent rigid-body refinement can deal with.  I'm
>> happy to give you more advice on this, off-line, because this kind of case
>> isn't something that we've figured out how to deal with automatically yet.
>> The optimal approach probably involves getting a deeper understanding of
>> commensurate modulation, which is another way of thinking about
>> pseudo-translations.
>>
>> Best wishes,
>>
>> Randy Read
>>
>> On 18 Nov 2013, at 09:19, #CHEN DAN#  wrote:
>>
>> Dear experts,
>>
>> I am working on one dataset (2.5A) which  was processed using space group
>> P43212 ( 107.9, 107.9, 313.7; 90, 90, 90).
>> After running MR with 6 molecules in ASU and one round of refmac, the R
>> factors are high (38%/45%).
>> I ran phenix.xtriage and found that translational pseudo symmetry is likely
>> present. It suggested that the space group is I4122 with the unit cell about
>> 1/3 smaller (I paste the patterson analyses below).
>> I tried to reprocess the data to get the suggested space group and unit cell
>> using HKL2000. But the index always gives a long c axis about 313A.
>> Could you provide any suggestions on how to proceed?
>>
>>  Patterson analyses
>> --
>>
>>  Largest Patterson peak with length larger than 15 Angstrom
>>
>>  Frac. coord.:0.5000.5000.174
>>  Distance to origin  :   93.757
>>  Height (origin=10

Re: [ccp4bb] sitcom

[Fabio Dall'Antonia]

Dear Tobias,

What exactly do you mean with "I can't get it to work" ? Do you get an 
error message or does the program simply compare the 100 SHELXD 
substructures against an ever-extending unique set (hence ignoring the 
command)?
I just checked myself with the Linux version 0.17.3 (which is the one 
you are using, I suppose) - it works in my hands. Using a thaumatin test 
structure, my input script looks like yours in principle, with 
"restrain_comp" in the last line (5th command).


To trace the error, please inspect your logfile:
1. Check that the "restrain comp" command appears in the "STDIN >>>" 
echo line
2. Check that in the recapitulation of recognized commands, "REFERENCE 
CHECK" is written as comparison mode.
3. If conditions 1. and 2. are true, the evaluation of substructures 
should just show the site position evaluation for every SHELXD trial, 
but not the extending "unique site" lists with the star-symbol (***) 
scoring for instances of agreements.


Are you comparing against a PDB file with a single substructure or 
against one with a complete, refined, protein structure? If the latter 
is the case, your read_sol line should read (e.g.):


read_sol REFI 1.0 name.pdb 18 SE

(an additional element argument appended to the command line)

I hope these comments can help you.
Kind regards,
Fabio



Subject: [ccp4bb] sitcom
Date: Mon, 18 Nov 2013 16:15:53 +0100
From: Tobias Weinert 
Reply-To: Tobias Weinert 
To: CCP4BB@JISCMAIL.AC.UK

Dear all,

i am trying to compare a set of solutions from shelxd against the 
correct solution and not against the generated “unique set" with 
sitcom my input file looks like this:


unit_cell XXX
space_group XXX
read_sol SOLUTION 1.0 sites.pdb 18
read_set LIST 0.1 try1_fa.lst 100 18

now i tried to add the restrain_comp keyword but i can’t get it to work.

thank you very much for your help,

Tobias






--
Dr. rer. nat. Fabio Dall'Antonia
European Molecular Biology Laboratory c/o DESY
Notkestraße 85, Bldg. 25a
D-22603 Hamburg

phone:  +49 (0)40 89902-178
fax:+49 (0)40 89902-149
e-mail: fabio.dallanto...@embl-hamburg.de

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Phil Evans]

I've generally found that adding lines to the "standard" table works, and they 
are not removed by editors


On 19 Nov 2013, at 09:32, Tim Gruene  wrote:

> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Dear Graeme,
> 
> On 11/19/2013 09:02 AM, Graeme Winter wrote:
>> [...] For the merged I/sig(I) Rpim is much more instructive. I'd
>> love it if people reported merged and unmerged I/sig(I), Rmerge,
>> Rmeas, Rpim, CC1/2, ... as each of these tells something
>> different.
> Depending on where you publish the editor will ask you to use their
> standard layout for the table which was probably last updated in the
> 1990's given the presence of something as sophisticated as an Rfree...
> 
> That's my recent experience, which undermined my preference for
> scientifically sound journals over tabloids. Unfortunately, it's the
> latter that funding agency like better ...
> 
> Best,
> Tim
> 
> 
>> 
>> Best wishes,
>> 
>> Graeme
>> 
>> Possibly useful papers:
>> 
>> http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html 
>> http://scripts.iucr.org/cgi-bin/paper?he0191 
>> http://scripts.iucr.org/cgi-bin/paper?he0268
>> 
>> 
>> 
>> 
>> On 19 November 2013 06:43, Shanti Pal Gangwar
>>  wrote:
>> 
>>> Dear  All
>>> 
>>> 
>>> Can anyone explain the meaning and relevance of data when the
>>> Rmerge is 100% in high resolution shell and I/sig(I) is 3.
>>> 
>>> 
>>> 
>>> Thanks
>>> 
>>> 
>>> 
>>> --  regards Shanti Pal Gangwar School of Life
>>> Sciences Jawaharlal Nehru University New Delhi-110067 India 
>>> Email:gangwar...@gmail.com
>>> 
>>> 
>>> 
>> 
> 
> - -- 
> - --
> Dr Tim Gruene
> Institut fuer anorganische Chemie
> Tammannstr. 4
> D-37077 Goettingen
> 
> GPG Key ID = A46BEE1A
> 
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.12 (GNU/Linux)
> Comment: Using GnuPG with Icedove - http://www.enigmail.net/
> 
> iD8DBQFSizAgUxlJ7aRr7hoRAtW8AJ9faxDJ6Wz2F5frob8PlOOXne2ZMACfdGxv
> fA0SSd2GsXKQRqZwg6MHjOk=
> =fyIi
> -END PGP SIGNATURE-

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Tim Gruene]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Dear Graeme,

On 11/19/2013 09:02 AM, Graeme Winter wrote:
> [...] For the merged I/sig(I) Rpim is much more instructive. I'd
> love it if people reported merged and unmerged I/sig(I), Rmerge,
> Rmeas, Rpim, CC1/2, ... as each of these tells something
> different.
Depending on where you publish the editor will ask you to use their
standard layout for the table which was probably last updated in the
1990's given the presence of something as sophisticated as an Rfree...

That's my recent experience, which undermined my preference for
scientifically sound journals over tabloids. Unfortunately, it's the
latter that funding agency like better ...

Best,
Tim


> 
> Best wishes,
> 
> Graeme
> 
> Possibly useful papers:
> 
> http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html 
> http://scripts.iucr.org/cgi-bin/paper?he0191 
> http://scripts.iucr.org/cgi-bin/paper?he0268
> 
> 
> 
> 
> On 19 November 2013 06:43, Shanti Pal Gangwar
>  wrote:
> 
>> Dear  All
>> 
>> 
>> Can anyone explain the meaning and relevance of data when the
>> Rmerge is 100% in high resolution shell and I/sig(I) is 3.
>> 
>> 
>> 
>> Thanks
>> 
>> 
>> 
>> --  regards Shanti Pal Gangwar School of Life
>> Sciences Jawaharlal Nehru University New Delhi-110067 India 
>> Email:gangwar...@gmail.com
>> 
>> 
>> 
> 

- -- 
- --
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen

GPG Key ID = A46BEE1A

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Icedove - http://www.enigmail.net/

iD8DBQFSizAgUxlJ7aRr7hoRAtW8AJ9faxDJ6Wz2F5frob8PlOOXne2ZMACfdGxv
fA0SSd2GsXKQRqZwg6MHjOk=
=fyIi
-END PGP SIGNATURE-

Re: [ccp4bb] Fix cell dimensions

[Tim Gruene]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Dear Niu,

in XDS the respective keywords in XDS.INP are "REFINE(IDXREF)",
"REFINE(INTEGRATE)" and "REFINE(CORRECT)".

If you do not want the cell to be refined, you have to set all three
keywords (i.e. ensure they are not commented out with a leading '!')
in XDS.INP and you must not mention the word 'CELL' on the line.

Best,
Tim

On 11/18/2013 10:48 PM, Niu Tou wrote:
> Dear All,
> 
> Does any one know how to strictly fix the cell dimensions during
> data processing? In HKL2000 there is only a keyword to define the
> longest vector. In XDS there is a option to input cell parameters,
> but sometimes the program would not follow the input values and
> switch back to the one it thinks best. Any suggestions will be 
> appreciated. Thanks!
> 
> Best, Niu
> 

- -- 
- --
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen

GPG Key ID = A46BEE1A

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Icedove - http://www.enigmail.net/

iD8DBQFSiy4+UxlJ7aRr7hoRAqzjAKCd0E6SBGmgDro+eTVXkmr1nFdpmgCgnCgv
FAVxVxvhKUSRkPTP63osrtQ=
=d5uB
-END PGP SIGNATURE-

Re: [ccp4bb] 100% Rmerge in high resolution shell

[Graeme Winter]

Usually this means that you have relatively high multiplicity, which
give-or-take improves the I/sig(I) by sqrt(m) where m is the multiplicity,
but also increases the Rmerge.

For any given narrow shell of reflections,

Rmerge ~ 0.8 / unmerged(I/sig(I))

merged(I/sig(I)) ~ sqrt(m) * unmerged(I/sig(I))

So it is perfectly possible to have unmerged I/sig(I) of 0.8 which will
give you an Rmerge of around 1.0, and have I/sig(I) (merged) around 3, by
having multiplciity 14 or so. I suggest that this is the case: if it is
much lower than this there is something odd going on.

For the merged I/sig(I) Rpim is much more instructive. I'd love it if
people reported merged and unmerged I/sig(I), Rmerge, Rmeas, Rpim, CC1/2,
... as each of these tells something different.

Best wishes,

Graeme

Possibly useful papers:

http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html
http://scripts.iucr.org/cgi-bin/paper?he0191
http://scripts.iucr.org/cgi-bin/paper?he0268




On 19 November 2013 06:43, Shanti Pal Gangwar  wrote:

> Dear  All
>
>
> Can anyone explain the meaning and relevance of data when the Rmerge is
> 100% in high resolution shell and I/sig(I) is 3.
>
>
>
> Thanks
>
>
>
> --
> 
> regards
> Shanti Pal Gangwar
> School of Life Sciences
> Jawaharlal Nehru University
> New Delhi-110067
> India
> Email:gangwar...@gmail.com
>
>
>