Re: [ccp4bb] Occupancy Refinement limitation

[Eleanor Dodson]

Low tech but simple.
Exclude both conformers from the refinement. (set occs to 0.00 so you still
see the coordinates in COOT)
Look at the difference map density and make a rough estimate of occupancy
from peak heights for a "fixed" bit of the conformers..

The relative ratios for different states should show up...
Eleanor

On Wed, 6 Sept 2023 at 11:49,  wrote:

> I would be tempted to try phenix.ensemble_refinement instead.
>
> Look at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3949522/
> or
> https://phenix-online.org/phenixwebsite_static/mainsite/files/presentations/ensemble_refinement_burnley_at_al_10DEC2012.pdf
>
> or https://elifesciences.org/articles/00311
>
> Tassos
>
> On 6 Sep 2023, at 12:37, Matt Mcleod  wrote:
>
> Hi,
>
> I should be a bit more specific.
>
> We have many crystal structures to indicate that a loop adopts
> conformation A and conformation B, or - the loop can be disordered where
> the electron density is washed out.  These states are dependent on how we
> perturb the system.
>
> We have a series of data, as a function of osmolyte concentration, that
> adjusts this equilibrium and we want to put a numerical value onto it.
> Resolution is between 1.7 - 2.0 A. This is non-anomalous data but certainly
> would be helpful in the future to include if possible some anomalous
> scatters. Have been using phenix.refine.
>
> So what we have done is modeled both conformation A and B and are refining
> the data in order to get the occupancies (fixing B-factors) regardless of
> if there is electron density present for one of the two conformations since
> in some cases (likely all) there will be a population of both A, B, and
> disordered and the relative true occupancies will move.  I am trying to
> sort out how accurate these occupancy values are as opposed to showing the
> electron density for each conformation fit at some common threshold.  My
> general sense is that if it is modeled but no electron density, there will
> be a non-zero value and vice versa if it is the only conformation present
> it will be less than unity.
>
> I will take a look at the references!  Very much appreciated.
> Matt
>
> On Wed, 6 Sept 2023 at 02:08, Eleanor Dodson <
> 176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
>
>> Well - occupancy refinement is particularly imprecise, and highly
>> correlated with temperature factors.
>> Also the population for a surface ARG or LYS may well have more than two
>> conformations, whereas some internal residue is better defined.
>> There is also the Q of solvent - dual occupancies will generate dual
>> solvent networks..
>>
>> ..You dont say what resolution your data are. Again at 0.8A you can be
>> confident - at 3A it is at best a guess. So I for one do not take the
>> numbers very seriously - they are a flag only.
>>
>> Presumably you didnt model the second conformation unless there was some
>> feature in an earlier map to suggest it existed?
>> Good luck Eleanor
>>
>>
>> On Wed, 6 Sept 2023 at 03:18, Pavel Afonine  wrote:
>>
>>> Hi Matt,
>>> I believe figure 3 here:
>>> https://www.nature.com/articles/s41467-018-06957-w
>>> is relevant to your question.
>>> Pavel
>>>
>>>
>>> On Tue, Sep 5, 2023 at 11:32 AM Matt McLeod 
>>> wrote:
>>>
 Hi all,

 I am trying to get some insight in the accuracy/precision of occupancy
 refinements.  I have done some 2-state occupancy refinements and have
 observed the refinement achieving ~0.25-0.3 occupancy for the minor
 population.  This population, when observing the electron density maps, had
 essentially no evidence for it being present.  I was wondering:

 What are the errors in the reported occupancies?

 Is there a lower and upper limit to occupancy refinements?  As in, if
 you occupancy refine two states and one is imaginary will it refine to
 approximately 1 and 0?  Or does the background noise always given a
 positive number to the imaginary set?  This would, to me at least, be the
 lower and upper limits to the occupancy refinements and could be used as a
 normalization factor for other atoms.  Maybe my logic is off...

 Any insight or literature would be appreciated!
 Matt

 

 To unsubscribe from the CCP4BB list, click the following link:
 https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

 This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a
 mailing list hosted by www.jiscmail.ac.uk, terms & conditions are
 available at https://www.jiscmail.ac.uk/policyandsecurity/

>>>
>>> --
>>>
>>> To unsubscribe from the CCP4BB list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>>>
>>
>> --
>>
>> To unsubscribe from the CCP4BB list, click the following link:
>> 

Re: [ccp4bb] Occupancy Refinement limitation

[a . perrakis]

I would be tempted to try phenix.ensemble_refinement instead.

Look at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3949522/
or 
https://phenix-online.org/phenixwebsite_static/mainsite/files/presentations/ensemble_refinement_burnley_at_al_10DEC2012.pdf
or https://elifesciences.org/articles/00311

Tassos

On 6 Sep 2023, at 12:37, Matt Mcleod 
mailto:mjmcleo...@gmail.com>> wrote:

Hi,

I should be a bit more specific.

We have many crystal structures to indicate that a loop adopts conformation A 
and conformation B, or - the loop can be disordered where the electron density 
is washed out.  These states are dependent on how we perturb the system.

We have a series of data, as a function of osmolyte concentration, that adjusts 
this equilibrium and we want to put a numerical value onto it.  Resolution is 
between 1.7 - 2.0 A. This is non-anomalous data but certainly would be helpful 
in the future to include if possible some anomalous scatters. Have been using 
phenix.refine.

So what we have done is modeled both conformation A and B and are refining the 
data in order to get the occupancies (fixing B-factors) regardless of if there 
is electron density present for one of the two conformations since in some 
cases (likely all) there will be a population of both A, B, and disordered and 
the relative true occupancies will move.  I am trying to sort out how accurate 
these occupancy values are as opposed to showing the electron density for each 
conformation fit at some common threshold.  My general sense is that if it is 
modeled but no electron density, there will be a non-zero value and vice versa 
if it is the only conformation present it will be less than unity.

I will take a look at the references!  Very much appreciated.
Matt

On Wed, 6 Sept 2023 at 02:08, Eleanor Dodson 
<176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk>
 wrote:
Well - occupancy refinement is particularly imprecise, and highly correlated 
with temperature factors.
Also the population for a surface ARG or LYS may well have more than two 
conformations, whereas some internal residue is better defined.
There is also the Q of solvent - dual occupancies will generate dual solvent 
networks..

..You dont say what resolution your data are. Again at 0.8A you can be 
confident - at 3A it is at best a guess. So I for one do not take the numbers 
very seriously - they are a flag only.

Presumably you didnt model the second conformation unless there was some 
feature in an earlier map to suggest it existed?
Good luck Eleanor


On Wed, 6 Sept 2023 at 03:18, Pavel Afonine 
mailto:pafon...@gmail.com>> wrote:
Hi Matt,
I believe figure 3 here:
https://www.nature.com/articles/s41467-018-06957-w
is relevant to your question.
Pavel


On Tue, Sep 5, 2023 at 11:32 AM Matt McLeod 
mailto:mjmcleo...@gmail.com>> wrote:
Hi all,

I am trying to get some insight in the accuracy/precision of occupancy 
refinements.  I have done some 2-state occupancy refinements and have observed 
the refinement achieving ~0.25-0.3 occupancy for the minor population.  This 
population, when observing the electron density maps, had essentially no 
evidence for it being present.  I was wondering:

What are the errors in the reported occupancies?

Is there a lower and upper limit to occupancy refinements?  As in, if you 
occupancy refine two states and one is imaginary will it refine to 
approximately 1 and 0?  Or does the background noise always given a positive 
number to the imaginary set?  This would, to me at least, be the lower and 
upper limits to the occupancy refinements and could be used as a normalization 
factor for other atoms.  Maybe my logic is off...

Any insight or literature would be appreciated!
Matt



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of 
www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions 
are available at https://www.jiscmail.ac.uk/policyandsecurity/



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1


--
Matthew Jordan McLeod, PhD
Post-Doctoral Fellow - Cornell University






To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message 

Re: [ccp4bb] Occupancy Refinement limitation

[Matt Mcleod]

Hi,

I should be a bit more specific.

We have many crystal structures to indicate that a loop adopts conformation
A and conformation B, or - the loop can be disordered where the electron
density is washed out.  These states are dependent on how we perturb the
system.

We have a series of data, as a function of osmolyte concentration, that
adjusts this equilibrium and we want to put a numerical value onto it.
Resolution is between 1.7 - 2.0 A. This is non-anomalous data but certainly
would be helpful in the future to include if possible some anomalous
scatters. Have been using phenix.refine.

So what we have done is modeled both conformation A and B and are refining
the data in order to get the occupancies (fixing B-factors) regardless of
if there is electron density present for one of the two conformations since
in some cases (likely all) there will be a population of both A, B, and
disordered and the relative true occupancies will move.  I am trying to
sort out how accurate these occupancy values are as opposed to showing the
electron density for each conformation fit at some common threshold.  My
general sense is that if it is modeled but no electron density, there will
be a non-zero value and vice versa if it is the only conformation present
it will be less than unity.

I will take a look at the references!  Very much appreciated.
Matt

On Wed, 6 Sept 2023 at 02:08, Eleanor Dodson <
176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:

> Well - occupancy refinement is particularly imprecise, and highly
> correlated with temperature factors.
> Also the population for a surface ARG or LYS may well have more than two
> conformations, whereas some internal residue is better defined.
> There is also the Q of solvent - dual occupancies will generate dual
> solvent networks..
>
> ..You dont say what resolution your data are. Again at 0.8A you can be
> confident - at 3A it is at best a guess. So I for one do not take the
> numbers very seriously - they are a flag only.
>
> Presumably you didnt model the second conformation unless there was some
> feature in an earlier map to suggest it existed?
> Good luck Eleanor
>
>
> On Wed, 6 Sept 2023 at 03:18, Pavel Afonine  wrote:
>
>> Hi Matt,
>> I believe figure 3 here:
>> https://www.nature.com/articles/s41467-018-06957-w
>> is relevant to your question.
>> Pavel
>>
>>
>> On Tue, Sep 5, 2023 at 11:32 AM Matt McLeod  wrote:
>>
>>> Hi all,
>>>
>>> I am trying to get some insight in the accuracy/precision of occupancy
>>> refinements.  I have done some 2-state occupancy refinements and have
>>> observed the refinement achieving ~0.25-0.3 occupancy for the minor
>>> population.  This population, when observing the electron density maps, had
>>> essentially no evidence for it being present.  I was wondering:
>>>
>>> What are the errors in the reported occupancies?
>>>
>>> Is there a lower and upper limit to occupancy refinements?  As in, if
>>> you occupancy refine two states and one is imaginary will it refine to
>>> approximately 1 and 0?  Or does the background noise always given a
>>> positive number to the imaginary set?  This would, to me at least, be the
>>> lower and upper limits to the occupancy refinements and could be used as a
>>> normalization factor for other atoms.  Maybe my logic is off...
>>>
>>> Any insight or literature would be appreciated!
>>> Matt
>>>
>>> 
>>>
>>> To unsubscribe from the CCP4BB list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>>>
>>> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a
>>> mailing list hosted by www.jiscmail.ac.uk, terms & conditions are
>>> available at https://www.jiscmail.ac.uk/policyandsecurity/
>>>
>>
>> --
>>
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>>
>
> --
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>


-- 
*Matthew Jordan McLeod, PhD*
*Post-Doctoral Fellow - Cornell University*



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] Occupancy Refinement limitation

[John R Helliwell]

Dear Matt,I believe this study we made may well assist your evaluations:-(IUCr) Experience with exchange and archiving of raw data: comparison of data from two diffractometers and four software packages on a series of lysozyme crystalsjournals.iucr.orgBest wishes,John Emeritus Professor John R Helliwell DScOn 5 Sep 2023, at 19:32, Matt McLeod  wrote:Hi all,I am trying to get some insight in the accuracy/precision of occupancy refinements.  I have done some 2-state occupancy refinements and have observed the refinement achieving ~0.25-0.3 occupancy for the minor population.  This population, when observing the electron density maps, had essentially no evidence for it being present.  I was wondering:What are the errors in the reported occupancies?Is there a lower and upper limit to occupancy refinements?  As in, if you occupancy refine two states and one is imaginary will it refine to approximately 1 and 0?  Or does the background noise always given a positive number to the imaginary set?  This would, to me at least, be the lower and upper limits to the occupancy refinements and could be used as a normalization factor for other atoms.  Maybe my logic is off...Any insight or literature would be appreciated!MattTo unsubscribe from the CCP4BB list, click the following link:https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/

To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

Re: [ccp4bb] Occupancy Refinement limitation

[Tim Gruene]

Hi Matt,

this can be done with SHELXL, even at 2-3A resolution. Anomalous data
do help reduce the correlation between B-factor and occupancy 
(see e.g. https://doi.org/10.1126/science.1254840 to stick with Pavel's
example)

SHELXL prints the correlation coefficient between occupancy and
b-factor, and also the ESD of the occupancy. The is more reliable with
least-square refinement, but for large structures, you may need a lot
of RAM.

PDB2INS (https://www.nature.com/articles/s41467-018-06957-w)
conveniently prepares the INS-file from your PDB for you.

Best wishes,
Tim

On Tue, 5 Sep 2023 19:31:25 +0100 Matt McLeod 
wrote:

> Hi all,
> 
> I am trying to get some insight in the accuracy/precision of
> occupancy refinements.  I have done some 2-state occupancy
> refinements and have observed the refinement achieving ~0.25-0.3
> occupancy for the minor population.  This population, when observing
> the electron density maps, had essentially no evidence for it being
> present.  I was wondering:
> 
> What are the errors in the reported occupancies?
> 
> Is there a lower and upper limit to occupancy refinements?  As in, if
> you occupancy refine two states and one is imaginary will it refine
> to approximately 1 and 0?  Or does the background noise always given
> a positive number to the imaginary set?  This would, to me at least,
> be the lower and upper limits to the occupancy refinements and could
> be used as a normalization factor for other atoms.  Maybe my logic is
> off...
> 
> Any insight or literature would be appreciated!
> Matt
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 
> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a
> mailing list hosted by www.jiscmail.ac.uk, terms & conditions are
> available at https://www.jiscmail.ac.uk/policyandsecurity/



-- 
--
Tim Gruene
Head of the Centre for X-ray Structure Analysis
Faculty of Chemistry
University of Vienna

Phone: +43-1-4277-70202

GPG Key ID = A46BEE1A



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/


pgp3Nciy9jGRf.pgp
Description: OpenPGP digital signature

Re: [ccp4bb] Occupancy Refinement limitation

[Eleanor Dodson]

Well - occupancy refinement is particularly imprecise, and highly
correlated with temperature factors.
Also the population for a surface ARG or LYS may well have more than two
conformations, whereas some internal residue is better defined.
There is also the Q of solvent - dual occupancies will generate dual
solvent networks..

..You dont say what resolution your data are. Again at 0.8A you can be
confident - at 3A it is at best a guess. So I for one do not take the
numbers very seriously - they are a flag only.

Presumably you didnt model the second conformation unless there was some
feature in an earlier map to suggest it existed?
Good luck Eleanor


On Wed, 6 Sept 2023 at 03:18, Pavel Afonine  wrote:

> Hi Matt,
> I believe figure 3 here:
> https://www.nature.com/articles/s41467-018-06957-w
> is relevant to your question.
> Pavel
>
>
> On Tue, Sep 5, 2023 at 11:32 AM Matt McLeod  wrote:
>
>> Hi all,
>>
>> I am trying to get some insight in the accuracy/precision of occupancy
>> refinements.  I have done some 2-state occupancy refinements and have
>> observed the refinement achieving ~0.25-0.3 occupancy for the minor
>> population.  This population, when observing the electron density maps, had
>> essentially no evidence for it being present.  I was wondering:
>>
>> What are the errors in the reported occupancies?
>>
>> Is there a lower and upper limit to occupancy refinements?  As in, if you
>> occupancy refine two states and one is imaginary will it refine to
>> approximately 1 and 0?  Or does the background noise always given a
>> positive number to the imaginary set?  This would, to me at least, be the
>> lower and upper limits to the occupancy refinements and could be used as a
>> normalization factor for other atoms.  Maybe my logic is off...
>>
>> Any insight or literature would be appreciated!
>> Matt
>>
>> 
>>
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>>
>> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a
>> mailing list hosted by www.jiscmail.ac.uk, terms & conditions are
>> available at https://www.jiscmail.ac.uk/policyandsecurity/
>>
>
> --
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] Occupancy Refinement limitation

[Pavel Afonine]

Hi Matt,
I believe figure 3 here:
https://www.nature.com/articles/s41467-018-06957-w
is relevant to your question.
Pavel


On Tue, Sep 5, 2023 at 11:32 AM Matt McLeod  wrote:

> Hi all,
>
> I am trying to get some insight in the accuracy/precision of occupancy
> refinements.  I have done some 2-state occupancy refinements and have
> observed the refinement achieving ~0.25-0.3 occupancy for the minor
> population.  This population, when observing the electron density maps, had
> essentially no evidence for it being present.  I was wondering:
>
> What are the errors in the reported occupancies?
>
> Is there a lower and upper limit to occupancy refinements?  As in, if you
> occupancy refine two states and one is imaginary will it refine to
> approximately 1 and 0?  Or does the background noise always given a
> positive number to the imaginary set?  This would, to me at least, be the
> lower and upper limits to the occupancy refinements and could be used as a
> normalization factor for other atoms.  Maybe my logic is off...
>
> Any insight or literature would be appreciated!
> Matt
>
> 
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>
> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a
> mailing list hosted by www.jiscmail.ac.uk, terms & conditions are
> available at https://www.jiscmail.ac.uk/policyandsecurity/
>



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Occupancy Refinement limitation

[Matt McLeod]

Hi all,

I am trying to get some insight in the accuracy/precision of occupancy 
refinements.  I have done some 2-state occupancy refinements and have observed 
the refinement achieving ~0.25-0.3 occupancy for the minor population.  This 
population, when observing the electron density maps, had essentially no 
evidence for it being present.  I was wondering:

What are the errors in the reported occupancies?

Is there a lower and upper limit to occupancy refinements?  As in, if you 
occupancy refine two states and one is imaginary will it refine to 
approximately 1 and 0?  Or does the background noise always given a positive 
number to the imaginary set?  This would, to me at least, be the lower and 
upper limits to the occupancy refinements and could be used as a normalization 
factor for other atoms.  Maybe my logic is off...

Any insight or literature would be appreciated!
Matt



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/