Re: [ccp4bb] Quote source inquiry

[Ian Tickle]

Hi Robbie

Yes I was (perhaps rashly!) assuming that the data are properly weighted
relative to the restraints.

Thanks for pointing that out.

Cheers

-- Ian


On Wed, 15 Jul 2020 at 20:23, Robbie Joosten 
wrote:

> Hi Ian,
>
> Errors in cell dimensions can have a large effect in MX with certain
> refinement doctrines. The school of "bond length rmsd must be $NUMBER"
> (which is still going strong unfortunately) will suffer from poor R-factors
> because the target cannot be satisfied without harming the fit to the data.
> At the same time if you have a a more relaxed approach to restraints than
> you might find systematic deviations in bond lengths. A test for that has
> been in WHAT_CHECK for decades and it actually works surprisingly well to
> detect cell dimension problems. That said, the problem is uncommon now.
>
> Cheers,
> Robbie
>
> > -Original Message-
> > From: CCP4 bulletin board  On Behalf Of Ian
> > Tickle
> > Sent: Wednesday, July 15, 2020 20:25
> > To: CCP4BB@JISCMAIL.AC.UK
> > Subject: Re: [ccp4bb] Quote source inquiry
> >
> >
> > Hi,
> >
> > There's one big difference between macromolecular and small molecule
> > refinement: except at ultra-high resolution the bond lengths in the
> former
> > are almost always strongly restrained, whereas those in the latter are
> almost
> > without exception completely unrestrained (except possibly bond lengths
> to
> > H atoms in XRD).  In other words, MX accurately determines the orthogonal
> > atomic co-ordinates, whereas XRD accurately determines their fractional
> co-
> > ordinates (it's no accident that the different programs output
> co-ordinates in
> > those formats).
> >
> > This means that since the cell dimensions are then used to convert
> fractional
> > to orthogonal in XRD, the final bond lengths will be much more sensitive
> to
> > errors in the cell dimensions, so having accurate cell dimensions is more
> > critical if you want accurate bond lengths (e.g. for use as restraints
> in MX!).
> > Obviously there's also a limit to the errors in the cell dimensions that
> can be
> > tolerated in MX: large errors will lead to errors in calculated d*
> values and
> > hence the scattering factors, which is likely to have a significant
> effect, and
> > there may be issues with VdW repulsions if the cell is too small (though
> it's
> > relatively easy for the structure to accommodate that).
> >
> > As Philip pointed out, the bond lengths will be totally insensitive to
> errors in
> > the uncertainties of the cell dimensions, whether artificially
> introduced or
> > poorly estimated from the data.  I don't know of any MX refinement
> > program (other than Shel-X) that takes the uncertainties in the cell
> > dimensions into account, even assuming that you have accurate values for
> > them.
> >
> > Also you should be careful not to confuse uncertainty (imprecision) with
> > error (inaccuracy).  The 'standard uncertainty' (s.u.) is the
> experimental
> > estimate of the 'standard deviation' (in the error)
> > (https://physics.nist.gov/cgi-bin/cuu/Info/Constants/definitions.html),
> and
> > the old term 'estimated standard deviation' (e.s.d.) was deprecated in
> 1993
> > (http://scripts.iucr.org/cgi-bin/paper?S0108767395002340).  You can have
> > an error in an uncertainty (which is what you were introducing in your
> test),
> > but you can't have an uncertainty in an error, since errors are by their
> nature
> > unknown anyway!
> >
> > It goes without saying that it's not a good idea to use bond lengths
> from a
> > restrained refinement as restraints in other refinements!
> >
> > Cheers
> >
> > -- Ian
> >
> >
> > On Wed, 15 Jul 2020 at 17:56, Jeffrey B Bonanno
> >  >  > wrote:
> >
> >
> >   Hi Phil,
> >
> >
> >
> >   Being young and impressionable, I only changed ZERR, and you are
> > quiet right the result is the rigorous and expected error propagation of
> > shelxl. Of course the more fun experiment would be in systematically
> > changing various values in UNIT to watch the molecule distort.
> >
> >
> >
> >   Hope all is well,
> >
> >   jbb
> >
> >
> >
> >   Jeffrey B. Bonanno, Ph.D.
> >
> >   Department of Biochemistry
> >
> >   Albert Einstein College of Medicine
> >
> >   1300 Morris Park Avenue
> >
> >   Bronx, NY 10461
> >
> >   off. 718-430-2452 fax. 718-430-8565
> >
> >   email jeffrey.bona...@einsteinmed.org
> > 
> >
> >
> >
> >   From: Jeffrey, Philip D. 
> >   Sent: Wednesday, July 15, 2020 12:47 PM
> >   To: CCP4BB@JISCMAIL.AC.UK  ;
> > Jeffrey B Bonanno  >  >
> >   Subject: Re: [ccp4bb] Quote source inquiry
> >
> >
> >
> > CAUTION: This email comes from an external source; the attachments and/or
> > links may compromise our secure environment. Do not open or click on
> > suspicious emails. Please click on the “Phish Alert” 

Re: [ccp4bb] Quote source inquiry

[Robbie Joosten]

Hi Ian,

Errors in cell dimensions can have a large effect in MX with certain refinement 
doctrines. The school of "bond length rmsd must be $NUMBER" (which is still 
going strong unfortunately) will suffer from poor R-factors because the target 
cannot be satisfied without harming the fit to the data. At the same time if 
you have a a more relaxed approach to restraints than you might find systematic 
deviations in bond lengths. A test for that has been in WHAT_CHECK for decades 
and it actually works surprisingly well to detect cell dimension problems. That 
said, the problem is uncommon now.

Cheers,
Robbie

> -Original Message-
> From: CCP4 bulletin board  On Behalf Of Ian
> Tickle
> Sent: Wednesday, July 15, 2020 20:25
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] Quote source inquiry
> 
> 
> Hi,
> 
> There's one big difference between macromolecular and small molecule
> refinement: except at ultra-high resolution the bond lengths in the former
> are almost always strongly restrained, whereas those in the latter are almost
> without exception completely unrestrained (except possibly bond lengths to
> H atoms in XRD).  In other words, MX accurately determines the orthogonal
> atomic co-ordinates, whereas XRD accurately determines their fractional co-
> ordinates (it's no accident that the different programs output co-ordinates in
> those formats).
> 
> This means that since the cell dimensions are then used to convert fractional
> to orthogonal in XRD, the final bond lengths will be much more sensitive to
> errors in the cell dimensions, so having accurate cell dimensions is more
> critical if you want accurate bond lengths (e.g. for use as restraints in 
> MX!).
> Obviously there's also a limit to the errors in the cell dimensions that can 
> be
> tolerated in MX: large errors will lead to errors in calculated d* values and
> hence the scattering factors, which is likely to have a significant effect, 
> and
> there may be issues with VdW repulsions if the cell is too small (though it's
> relatively easy for the structure to accommodate that).
> 
> As Philip pointed out, the bond lengths will be totally insensitive to errors 
> in
> the uncertainties of the cell dimensions, whether artificially introduced or
> poorly estimated from the data.  I don't know of any MX refinement
> program (other than Shel-X) that takes the uncertainties in the cell
> dimensions into account, even assuming that you have accurate values for
> them.
> 
> Also you should be careful not to confuse uncertainty (imprecision) with
> error (inaccuracy).  The 'standard uncertainty' (s.u.) is the experimental
> estimate of the 'standard deviation' (in the error)
> (https://physics.nist.gov/cgi-bin/cuu/Info/Constants/definitions.html), and
> the old term 'estimated standard deviation' (e.s.d.) was deprecated in 1993
> (http://scripts.iucr.org/cgi-bin/paper?S0108767395002340).  You can have
> an error in an uncertainty (which is what you were introducing in your test),
> but you can't have an uncertainty in an error, since errors are by their 
> nature
> unknown anyway!
> 
> It goes without saying that it's not a good idea to use bond lengths from a
> restrained refinement as restraints in other refinements!
> 
> Cheers
> 
> -- Ian
> 
> 
> On Wed, 15 Jul 2020 at 17:56, Jeffrey B Bonanno
>   > wrote:
> 
> 
>   Hi Phil,
> 
> 
> 
>   Being young and impressionable, I only changed ZERR, and you are
> quiet right the result is the rigorous and expected error propagation of
> shelxl. Of course the more fun experiment would be in systematically
> changing various values in UNIT to watch the molecule distort.
> 
> 
> 
>   Hope all is well,
> 
>   jbb
> 
> 
> 
>   Jeffrey B. Bonanno, Ph.D.
> 
>   Department of Biochemistry
> 
>   Albert Einstein College of Medicine
> 
>   1300 Morris Park Avenue
> 
>   Bronx, NY 10461
> 
>   off. 718-430-2452 fax. 718-430-8565
> 
>   email jeffrey.bona...@einsteinmed.org
> 
> 
> 
> 
>   From: Jeffrey, Philip D. 
>   Sent: Wednesday, July 15, 2020 12:47 PM
>   To: CCP4BB@JISCMAIL.AC.UK  ;
> Jeffrey B Bonanno   >
>   Subject: Re: [ccp4bb] Quote source inquiry
> 
> 
> 
> CAUTION: This email comes from an external source; the attachments and/or
> links may compromise our secure environment. Do not open or click on
> suspicious emails. Please click on the “Phish Alert” button on the top right 
> of
> the Outlook dashboard to report any suspicious emails.
> 
>   :: took a working dataset and increased (only) the error on unit cell
> dimensions in the instruction file for the final round of full matrix :: least
> squares refinement in shelxl. Sure enough, the errors on the bonds and
> angles shot up. I was more careful
> 
> 
> 
>   Question: did you change the unit cell dimensions (UNIT) 

Re: [ccp4bb] Quote source inquiry

[Jon Cooper]

Hello,

being a pedant, I guess you mean 'cell':

http://xray.chem.ualberta.ca/xray/shelxl/cell.htm

rather than 'unit':

http://xray.chem.ualberta.ca/xray/shelxl/UNIT.htm

;-?

Best wishes, Jon Cooper

 Original Message 
On 15 Jul 2020, 17:55, Jeffrey B Bonanno wrote:

> Hi Phil,
>
> Being young and impressionable, I only changed ZERR, and you are quiet right 
> the result is the rigorous and expected error propagation of shelxl. Of 
> course the more fun experiment would be in systematically changing various 
> values in UNIT to watch the molecule distort.
>
> Hope all is well,
> jbb
>
> Jeffrey B. Bonanno, Ph.D.
> Department of Biochemistry
> Albert Einstein College of Medicine
> 1300 Morris Park Avenue
> Bronx, NY 10461
> off. 718-430-2452 fax. 718-430-8565
> email jeffrey.bona...@einsteinmed.org
>
> From: Jeffrey, Philip D. 
> Sent: Wednesday, July 15, 2020 12:47 PM
> To: CCP4BB@JISCMAIL.AC.UK; Jeffrey B Bonanno 
> Subject: Re: [ccp4bb] Quote source inquiry
>
> CAUTION: This email comes from an external source; the attachments and/or 
> links may compromise our secure environment. Do not open or click on 
> suspicious emails. Please click on the “Phish Alert” button on the top right 
> of the Outlook dashboard to report any suspicious emails.
>
> :: took a working dataset and increased (only) the error on unit cell 
> dimensions in the instruction file for the final round of full matrix :: 
> least squares refinement in shelxl. Sure enough, the errors on the bonds and 
> angles shot up. I was more careful
>
> Question: did you change the unit cell dimensions (UNIT) or the reported 
> standard error in the unit cell dimensions (ZERR) ? If just the latter don't 
> you think that the error propagation is just a factor of SHELXL converting 
> from fractional to orthogonal coordinates to give you bond lengths and bond 
> angles (i.e. the bonds and angles would be numerically the same, but the 
> estimated error associated with them would be higher). Did the e.s.d.'s of 
> the actual coordinates in fractional space change ?
>
> Phil Jeffrey
>
> Princeton
>
> ---
>
> From: CCP4 bulletin board  on behalf of Jeffrey B 
> Bonanno 
> Sent: Wednesday, July 15, 2020 12:36 PM
> To: CCP4BB@JISCMAIL.AC.UK 
> Subject: Re: [ccp4bb] Quote source inquiry
>
> Hi Gerard and Bernhard,
>
> As a postdoc in an unnamed small molecule lab, I was instructed by my lab 
> head to get better unit cell estimates prior to data collection owing to 
> error propagation from the uncertainty on cell dimensions through to the esd 
> on atomic bond lengths and angles when refining in shelxl. To verify this 
> (what, you believed everything your postdoc advisor told you?), I took a 
> working dataset and increased (only) the error on unit cell dimensions in the 
> instruction file for the final round of full matrix least squares refinement 
> in shelxl. Sure enough, the errors on the bonds and angles shot up. I was 
> more careful in determining the unit cell thereafter. That is, until, I 
> became a macromolecular crystallographer...
>
> After an inciteful (sp? lol) discussion with Wladek about cell dimensions, I 
> was directed to read this paper:
>
> Acta Crystallogr D Biol Crystallogr. 2015 Nov 1; 71(Pt 11): 2217–2226.
>
> Have a look, it is interesting.
>
> Having never followed up on these studies to see what happened to bonds and 
> angles in proteins and their ligands when varying cell dimensions, I can't 
> say with any confidence. However, I would guess that the quality of the 
> refined ligand coordinates could only be as good as some combination of 
> factors including but not limited to 1) the data (resolution, B factor, etc), 
> 2) the actual occupancy of the ligand, and 3) the restraints employed.
>
> jbb
>
> Jeffrey B. Bonanno, Ph.D.
> Department of Biochemistry
> Albert Einstein College of Medicine
> 1300 Morris Park Avenue
> Bronx, NY 10461
> off. 718-430-2452 fax. 718-430-8565
> email jeffrey.bona...@einsteinmed.org
>
> -Original Message-
> From: CCP4 bulletin board  On Behalf Of Gerard DVD 
> Kleywegt
> Sent: Wednesday, July 15, 2020 11:49 AM
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] Quote source inquiry
>
> Well, I've had this in my CSHL X-ray Course talk for many years.
>
> In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray 
> crystallography is a notoriously poor method for determining the structure of 
> small molecules that are bound to macromolecules [...]" and then goes on to 
> explain why this is the case.
>
> In the attached 2003 paper (pooling the wisdom of several of the usual 
> suspects, including Eleanor) it says something similar (p 1057):
>
> "Coordinates of molecules that have been determined in complex with 
> macromolecules previously can of course also be retrieved from the PDB 
> (Bernstein et al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones, 
> 1998), or CHEMPDB (Boutselakis 

Re: [ccp4bb] Quote source inquiry

[Ian Tickle]

Hi,

There's one big difference between macromolecular and small molecule
refinement: except at ultra-high resolution the bond lengths in the former
are almost always strongly restrained, whereas those in the latter are
almost without exception completely unrestrained (except possibly bond
lengths to H atoms in XRD).  In other words, MX accurately determines the
orthogonal atomic co-ordinates, whereas XRD accurately determines their
fractional co-ordinates (it's no accident that the different programs
output co-ordinates in those formats).

This means that since the cell dimensions are then used to convert
fractional to orthogonal in XRD, the final bond lengths will be much more
sensitive to errors in the cell dimensions, so having accurate cell
dimensions is more critical if you want accurate bond lengths (e.g. for use
as restraints in MX!).  Obviously there's also a limit to the errors in the
cell dimensions that can be tolerated in MX: large errors will lead to
errors in calculated d* values and hence the scattering factors, which is
likely to have a significant effect, and there may be issues with VdW
repulsions if the cell is too small (though it's relatively easy for the
structure to accommodate that).

As Philip pointed out, the bond lengths will be totally insensitive to
errors in the uncertainties of the cell dimensions, whether artificially
introduced or poorly estimated from the data.  I don't know of any MX
refinement program (other than Shel-X) that takes the uncertainties in the
cell dimensions into account, even assuming that you have accurate values
for them.

Also you should be careful not to confuse uncertainty (imprecision) with
error (inaccuracy).  The 'standard uncertainty' (s.u.) is the experimental
estimate of the 'standard deviation' (in the error) (
https://physics.nist.gov/cgi-bin/cuu/Info/Constants/definitions.html), and
the old term 'estimated standard deviation' (e.s.d.) was deprecated in 1993
(http://scripts.iucr.org/cgi-bin/paper?S0108767395002340).  You can have an
error in an uncertainty (which is what you were introducing in your test),
but you can't have an uncertainty in an error, since errors are by their
nature unknown anyway!

It goes without saying that it's not a good idea to use bond lengths from a
restrained refinement as restraints in other refinements!

Cheers

-- Ian


On Wed, 15 Jul 2020 at 17:56, Jeffrey B Bonanno <
jeffrey.bona...@einsteinmed.org> wrote:

> Hi Phil,
>
>
>
> Being young and impressionable, I only changed ZERR, and you are quiet
> right the result is the rigorous and expected error propagation of shelxl.
> Of course the more fun experiment would be in systematically changing
> various values in UNIT to watch the molecule distort.
>
>
>
> Hope all is well,
>
> jbb
>
>
>
> Jeffrey B. Bonanno, Ph.D.
>
> Department of Biochemistry
>
> Albert Einstein College of Medicine
>
> 1300 Morris Park Avenue
>
> Bronx, NY 10461
>
> off. 718-430-2452 fax. 718-430-8565
>
> email jeffrey.bona...@einsteinmed.org
>
>
>
> *From:* Jeffrey, Philip D. 
> *Sent:* Wednesday, July 15, 2020 12:47 PM
> *To:* CCP4BB@JISCMAIL.AC.UK; Jeffrey B Bonanno <
> jeffrey.bona...@einsteinmed.org>
> *Subject:* Re: [ccp4bb] Quote source inquiry
>
>
>
> *CAUTION: This email comes from an external source; the attachments and/or
> links may compromise our secure environment. Do not open or click on
> suspicious emails. Please click on the “Phish Alert” button on the top
> right of the Outlook dashboard to report any suspicious emails.*
>
> :: took a working dataset and increased (only) the error on unit cell
> dimensions in the instruction file for the final round of full matrix ::
> least squares refinement in shelxl. Sure enough, the errors on the bonds
> and angles shot up. I was more careful
>
>
>
> Question: did you change the unit cell dimensions (UNIT) or the reported
> standard error in the unit cell dimensions (ZERR) ?  If just the latter
> don't you think that the error propagation is just a factor of SHELXL
> converting from fractional to orthogonal coordinates to give you bond
> lengths and bond angles (i.e. the bonds and angles would be numerically the
> same, but the estimated error associated with them would be higher).  Did
> the e.s.d.'s of the actual coordinates in fractional space change ?
>
>
>
> Phil Jeffrey
>
> Princeton
> --
>
> *From:* CCP4 bulletin board  on behalf of Jeffrey
> B Bonanno 
> *Sent:* Wednesday, July 15, 2020 12:36 PM
> *To:* CCP4BB@JISCMAIL.AC.UK 
> *Subject:* Re: [ccp4bb] Quote source inquiry
>
>
>
> Hi Gerard and Bernhard,
>
> As a postdoc in an unnamed small molecule lab, I was instructed by my lab
> head to get better unit cell estimates prior to data collection owing to
> error propagation from the uncertainty on cell dimensions through to the
> esd on atomic bond lengths and angles when refining in shelxl. To verify
> this (what, you believed everything your postdoc advisor told you?), I took
> a working dataset 

Re: [ccp4bb] Quote source inquiry

[Jeffrey B Bonanno]

Hi Phil,

Being young and impressionable, I only changed ZERR, and you are quiet right 
the result is the rigorous and expected error propagation of shelxl. Of course 
the more fun experiment would be in systematically changing various values in 
UNIT to watch the molecule distort.

Hope all is well,
jbb

Jeffrey B. Bonanno, Ph.D.
Department of Biochemistry
Albert Einstein College of Medicine
1300 Morris Park Avenue
Bronx, NY 10461
off. 718-430-2452 fax. 718-430-8565
email jeffrey.bona...@einsteinmed.org

From: Jeffrey, Philip D. 
Sent: Wednesday, July 15, 2020 12:47 PM
To: CCP4BB@JISCMAIL.AC.UK; Jeffrey B Bonanno 
Subject: Re: [ccp4bb] Quote source inquiry

CAUTION: This email comes from an external source; the attachments and/or links 
may compromise our secure environment. Do not open or click on suspicious 
emails. Please click on the "Phish Alert" button on the top right of the 
Outlook dashboard to report any suspicious emails.
:: took a working dataset and increased (only) the error on unit cell 
dimensions in the instruction file for the final round of full matrix :: least 
squares refinement in shelxl. Sure enough, the errors on the bonds and angles 
shot up. I was more careful

Question: did you change the unit cell dimensions (UNIT) or the reported 
standard error in the unit cell dimensions (ZERR) ?  If just the latter don't 
you think that the error propagation is just a factor of SHELXL converting from 
fractional to orthogonal coordinates to give you bond lengths and bond angles 
(i.e. the bonds and angles would be numerically the same, but the estimated 
error associated with them would be higher).  Did the e.s.d.'s of the actual 
coordinates in fractional space change ?

Phil Jeffrey
Princeton

From: CCP4 bulletin board mailto:CCP4BB@JISCMAIL.AC.UK>> 
on behalf of Jeffrey B Bonanno 
mailto:jeffrey.bona...@einsteinmed.org>>
Sent: Wednesday, July 15, 2020 12:36 PM
To: CCP4BB@JISCMAIL.AC.UK 
mailto:CCP4BB@JISCMAIL.AC.UK>>
Subject: Re: [ccp4bb] Quote source inquiry

Hi Gerard and Bernhard,

As a postdoc in an unnamed small molecule lab, I was instructed by my lab head 
to get better unit cell estimates prior to data collection owing to error 
propagation from the uncertainty on cell dimensions through to the esd on 
atomic bond lengths and angles when refining in shelxl. To verify this (what, 
you believed everything your postdoc advisor told you?), I took a working 
dataset and increased (only) the error on unit cell dimensions in the 
instruction file for the final round of full matrix least squares refinement in 
shelxl. Sure enough, the errors on the bonds and angles shot up. I was more 
careful in determining the unit cell thereafter. That is, until, I became a 
macromolecular crystallographer...

After an inciteful (sp? lol) discussion with Wladek about cell dimensions, I 
was directed to read this paper:

Acta Crystallogr D Biol Crystallogr. 2015 Nov 1; 71(Pt 11): 2217-2226.

Have a look, it is interesting.

Having never followed up on these studies to see what happened to bonds and 
angles in proteins and their ligands when varying cell dimensions, I can't say 
with any confidence. However, I would guess that the quality of the refined 
ligand coordinates could only be as good as some combination of factors 
including but not limited to 1) the data (resolution, B factor, etc), 2) the 
actual occupancy of the ligand, and 3) the restraints employed.

jbb

Jeffrey B. Bonanno, Ph.D.
Department of Biochemistry
Albert Einstein College of Medicine
1300 Morris Park Avenue
Bronx, NY 10461
off. 718-430-2452 fax. 718-430-8565
email jeffrey.bona...@einsteinmed.org


-Original Message-
From: CCP4 bulletin board mailto:CCP4BB@JISCMAIL.AC.UK>> 
On Behalf Of Gerard DVD Kleywegt
Sent: Wednesday, July 15, 2020 11:49 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Quote source inquiry

Well, I've had this in my CSHL X-ray Course talk for many years.

In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray 
crystallography is a notoriously poor method for determining the structure of 
small molecules that are bound to macromolecules [...]" and then goes on to 
explain why this is the case.

In the attached 2003 paper (pooling the wisdom of several of the usual 
suspects, including Eleanor) it says something similar (p 1057):

"Coordinates of molecules that have been determined in complex with 
macromolecules previously can of course also be retrieved from the PDB 
(Bernstein et al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones, 
1998), or CHEMPDB (Boutselakis et al., 2003). However, one should keep in mind 
that these coordinates are the result of refinement against comparatively
low-resolu- tion data where the small molecule constituted only a minute 
fraction of the total scattering matter. This makes these coordinates 
inherently much 

Re: [ccp4bb] Quote source inquiry

[Jeffrey, Philip D.]

:: took a working dataset and increased (only) the error on unit cell 
dimensions in the instruction file for the final round of full matrix :: least 
squares refinement in shelxl. Sure enough, the errors on the bonds and angles 
shot up. I was more careful

Question: did you change the unit cell dimensions (UNIT) or the reported 
standard error in the unit cell dimensions (ZERR) ?  If just the latter don't 
you think that the error propagation is just a factor of SHELXL converting from 
fractional to orthogonal coordinates to give you bond lengths and bond angles 
(i.e. the bonds and angles would be numerically the same, but the estimated 
error associated with them would be higher).  Did the e.s.d.'s of the actual 
coordinates in fractional space change ?

Phil Jeffrey
Princeton

From: CCP4 bulletin board  on behalf of Jeffrey B 
Bonanno 
Sent: Wednesday, July 15, 2020 12:36 PM
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] Quote source inquiry

Hi Gerard and Bernhard,

As a postdoc in an unnamed small molecule lab, I was instructed by my lab head 
to get better unit cell estimates prior to data collection owing to error 
propagation from the uncertainty on cell dimensions through to the esd on 
atomic bond lengths and angles when refining in shelxl. To verify this (what, 
you believed everything your postdoc advisor told you?), I took a working 
dataset and increased (only) the error on unit cell dimensions in the 
instruction file for the final round of full matrix least squares refinement in 
shelxl. Sure enough, the errors on the bonds and angles shot up. I was more 
careful in determining the unit cell thereafter. That is, until, I became a 
macromolecular crystallographer...

After an inciteful (sp? lol) discussion with Wladek about cell dimensions, I 
was directed to read this paper:

Acta Crystallogr D Biol Crystallogr. 2015 Nov 1; 71(Pt 11): 2217–2226.

Have a look, it is interesting.

Having never followed up on these studies to see what happened to bonds and 
angles in proteins and their ligands when varying cell dimensions, I can't say 
with any confidence. However, I would guess that the quality of the refined 
ligand coordinates could only be as good as some combination of factors 
including but not limited to 1) the data (resolution, B factor, etc), 2) the 
actual occupancy of the ligand, and 3) the restraints employed.

jbb

Jeffrey B. Bonanno, Ph.D.
Department of Biochemistry
Albert Einstein College of Medicine
1300 Morris Park Avenue
Bronx, NY 10461
off. 718-430-2452 fax. 718-430-8565
email jeffrey.bona...@einsteinmed.org


-Original Message-
From: CCP4 bulletin board  On Behalf Of Gerard DVD 
Kleywegt
Sent: Wednesday, July 15, 2020 11:49 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Quote source inquiry

Well, I've had this in my CSHL X-ray Course talk for many years.

In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray 
crystallography is a notoriously poor method for determining the structure of 
small molecules that are bound to macromolecules [...]" and then goes on to 
explain why this is the case.

In the attached 2003 paper (pooling the wisdom of several of the usual 
suspects, including Eleanor) it says something similar (p 1057):

"Coordinates of molecules that have been determined in complex with 
macromolecules previously can of course also be retrieved from the PDB 
(Bernstein et al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones, 
1998), or CHEMPDB (Boutselakis et al., 2003). However, one should keep in mind 
that these coordinates are the result of refinement against comparatively
low-resolu- tion data where the small molecule constituted only a minute 
fraction of the total scattering matter. This makes these coordinates 
inherently much less accurate than those obtained from the CSD. In addition, 
the coordi- nates may contain errors due to the use of incorrect restraints.
Hence, such coordinate sets should only be used as a last resort, and only 
after verification that they are reliable. The latter can be facilitated by 
inspection of the electron density for the compound in question, for instance 
at the Uppsala Electron-Density Server (http:// fsrv1.bmc.uu.se/eds) (G.J.K.
et al., submitted)."

Happy to be confused with George though!

--Gerard (no, the other one)



On Tue, 14 Jul 2020, Bernhard Rupp wrote:

> Hi Fellows,
>
>
>
> afaicrimps (as far as I can recall in my progressing senility)
> someone once wrote/stated/cursed somewhere that "Macromolecular
> refinement is not a small molecule structure determination method".
>
>
>
> Any citable source - George Sheldrick might be a suspect.
>
>
>
> Thanks & best regards, BR
>
>
>
> --
>
> Bernhard Rupp
>
>  hofkristallamt.org%2Fdata=02%7C01%7Cjeffrey.bonanno%40einsteinmed
> 

Re: [ccp4bb] Quote source inquiry

[Gerard DVD Kleywegt]

Thanks - very interesting paper, Jeffrey.

I think your analysis is correct, but you forgot: (4) the skill of the 
crystallographer...


Best wishes,

--Gerard


On Wed, 15 Jul 2020, Jeffrey B Bonanno wrote:


Hi Gerard and Bernhard,

As a postdoc in an unnamed small molecule lab, I was instructed by my lab head 
to get better unit cell estimates prior to data collection owing to error 
propagation from the uncertainty on cell dimensions through to the esd on 
atomic bond lengths and angles when refining in shelxl. To verify this (what, 
you believed everything your postdoc advisor told you?), I took a working 
dataset and increased (only) the error on unit cell dimensions in the 
instruction file for the final round of full matrix least squares refinement in 
shelxl. Sure enough, the errors on the bonds and angles shot up. I was more 
careful in determining the unit cell thereafter. That is, until, I became a 
macromolecular crystallographer...

After an inciteful (sp? lol) discussion with Wladek about cell dimensions, I 
was directed to read this paper:

Acta Crystallogr D Biol Crystallogr. 2015 Nov 1; 71(Pt 11): 2217–2226.

Have a look, it is interesting.

Having never followed up on these studies to see what happened to bonds and 
angles in proteins and their ligands when varying cell dimensions, I can't 
say with any confidence. However, I would guess that the quality of the 
refined ligand coordinates could only be as good as some combination of 
factors including but not limited to 1) the data (resolution, B factor, 
etc), 2) the actual occupancy of the ligand, and 3) the restraints employed.


jbb

Jeffrey B. Bonanno, Ph.D.
Department of Biochemistry
Albert Einstein College of Medicine
1300 Morris Park Avenue
Bronx, NY 10461
off. 718-430-2452 fax. 718-430-8565
email jeffrey.bona...@einsteinmed.org


-Original Message-
From: CCP4 bulletin board  On Behalf Of Gerard DVD 
Kleywegt
Sent: Wednesday, July 15, 2020 11:49 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Quote source inquiry

Well, I've had this in my CSHL X-ray Course talk for many years.

In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray 
crystallography is a notoriously poor method for determining the structure of small 
molecules that are bound to macromolecules [...]" and then goes on to explain why 
this is the case.

In the attached 2003 paper (pooling the wisdom of several of the usual 
suspects, including Eleanor) it says something similar (p 1057):

"Coordinates of molecules that have been determined in complex with 
macromolecules previously can of course also be retrieved from the PDB (Bernstein et 
al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones, 1998), or CHEMPDB 
(Boutselakis et al., 2003). However, one should keep in mind that these coordinates 
are the result of refinement against comparatively
low-resolu- tion data where the small molecule constituted only a minute 
fraction of the total scattering matter. This makes these coordinates 
inherently much less accurate than those obtained from the CSD. In addition, 
the coordi- nates may contain errors due to the use of incorrect restraints.
Hence, such coordinate sets should only be used as a last resort, and only 
after verification that they are reliable. The latter can be facilitated by 
inspection of the electron density for the compound in question, for instance 
at the Uppsala Electron-Density Server (http:// fsrv1.bmc.uu.se/eds) (G.J.K.
et al., submitted)."

Happy to be confused with George though!

--Gerard (no, the other one)



On Tue, 14 Jul 2020, Bernhard Rupp wrote:


Hi Fellows,



afaicrimps (as far as I can recall in my progressing senility)
someone once wrote/stated/cursed somewhere that "Macromolecular
refinement is not a small molecule structure determination method".



Any citable source - George Sheldrick might be a suspect.



Thanks & best regards, BR



--

Bernhard Rupp


https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.h
ofkristallamt.org%2Fdata=02%7C01%7Cjeffrey.bonanno%40einsteinmed.
org%7Cf0b1878c0df040bb2e5308d828d6b0ed%7C9c01f0fd65e040c089a82dfd51e62
025%7C0%7C0%7C637304250594203532sdata=9FFCpYd9D%2BrHR48DtL%2BglMW
2dbZ%2FmSTw0fF88XRwR7o%3Dreserved=0

 b...@hofkristallamt.org

+1 925 209 7429

+43 676 571 0536

--

Many plausible ideas vanish

at the presence of thought

--




##
##

To unsubscribe from the CCP4BB list, 

Re: [ccp4bb] Quote source inquiry

[Jeffrey B Bonanno]

Hi Gerard and Bernhard,

As a postdoc in an unnamed small molecule lab, I was instructed by my lab head 
to get better unit cell estimates prior to data collection owing to error 
propagation from the uncertainty on cell dimensions through to the esd on 
atomic bond lengths and angles when refining in shelxl. To verify this (what, 
you believed everything your postdoc advisor told you?), I took a working 
dataset and increased (only) the error on unit cell dimensions in the 
instruction file for the final round of full matrix least squares refinement in 
shelxl. Sure enough, the errors on the bonds and angles shot up. I was more 
careful in determining the unit cell thereafter. That is, until, I became a 
macromolecular crystallographer...

After an inciteful (sp? lol) discussion with Wladek about cell dimensions, I 
was directed to read this paper:

Acta Crystallogr D Biol Crystallogr. 2015 Nov 1; 71(Pt 11): 2217–2226.

Have a look, it is interesting.

Having never followed up on these studies to see what happened to bonds and 
angles in proteins and their ligands when varying cell dimensions, I can't say 
with any confidence. However, I would guess that the quality of the refined 
ligand coordinates could only be as good as some combination of factors 
including but not limited to 1) the data (resolution, B factor, etc), 2) the 
actual occupancy of the ligand, and 3) the restraints employed.

jbb

Jeffrey B. Bonanno, Ph.D.
Department of Biochemistry
Albert Einstein College of Medicine
1300 Morris Park Avenue
Bronx, NY 10461
off. 718-430-2452 fax. 718-430-8565
email jeffrey.bona...@einsteinmed.org


-Original Message-
From: CCP4 bulletin board  On Behalf Of Gerard DVD 
Kleywegt
Sent: Wednesday, July 15, 2020 11:49 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Quote source inquiry

Well, I've had this in my CSHL X-ray Course talk for many years.

In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray 
crystallography is a notoriously poor method for determining the structure of 
small molecules that are bound to macromolecules [...]" and then goes on to 
explain why this is the case.

In the attached 2003 paper (pooling the wisdom of several of the usual 
suspects, including Eleanor) it says something similar (p 1057):

"Coordinates of molecules that have been determined in complex with 
macromolecules previously can of course also be retrieved from the PDB 
(Bernstein et al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones, 
1998), or CHEMPDB (Boutselakis et al., 2003). However, one should keep in mind 
that these coordinates are the result of refinement against comparatively
low-resolu- tion data where the small molecule constituted only a minute 
fraction of the total scattering matter. This makes these coordinates 
inherently much less accurate than those obtained from the CSD. In addition, 
the coordi- nates may contain errors due to the use of incorrect restraints.
Hence, such coordinate sets should only be used as a last resort, and only 
after verification that they are reliable. The latter can be facilitated by 
inspection of the electron density for the compound in question, for instance 
at the Uppsala Electron-Density Server (http:// fsrv1.bmc.uu.se/eds) (G.J.K.
et al., submitted)."

Happy to be confused with George though!

--Gerard (no, the other one)



On Tue, 14 Jul 2020, Bernhard Rupp wrote:

> Hi Fellows,
>
>
>
> afaicrimps (as far as I can recall in my progressing senility)  
> someone once wrote/stated/cursed somewhere that "Macromolecular 
> refinement is not a small molecule structure determination method".
>
>
>
> Any citable source - George Sheldrick might be a suspect.
>
>
>
> Thanks & best regards, BR
>
>
>
> --
>
> Bernhard Rupp
>
>  hofkristallamt.org%2Fdata=02%7C01%7Cjeffrey.bonanno%40einsteinmed
> .org%7Cf0b1878c0df040bb2e5308d828d6b0ed%7C9c01f0fd65e040c089a82dfd51e6
> 2025%7C0%7C0%7C637304250594203532sdata=9FFCpYd9D%2BrHR48DtL%2BglM
> W2dbZ%2FmSTw0fF88XRwR7o%3Dreserved=0> 
> https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.h
> ofkristallamt.org%2Fdata=02%7C01%7Cjeffrey.bonanno%40einsteinmed.
> org%7Cf0b1878c0df040bb2e5308d828d6b0ed%7C9c01f0fd65e040c089a82dfd51e62
> 025%7C0%7C0%7C637304250594203532sdata=9FFCpYd9D%2BrHR48DtL%2BglMW
> 2dbZ%2FmSTw0fF88XRwR7o%3Dreserved=0
>
>  b...@hofkristallamt.org
>
> +1 925 209 7429
>
> +43 676 571 0536
>
> --
>
> Many plausible ideas vanish
>
> at the presence of thought
>
> --
>
>
>
>
> ##
> ##
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.
> 

Re: [ccp4bb] Quote source inquiry

[Bernhard Rupp]

Close enough verbatim and dead on in spirit.
Many thanks, BR

-Original Message-
From: Gerard DVD Kleywegt  
Sent: Wednesday, July 15, 2020 08:49
To: b...@hofkristallamt.org
Cc: CCP4 Bulletin Board 
Subject: Re: [ccp4bb] Quote source inquiry

Well, I've had this in my CSHL X-ray Course talk for many years.

In the attached 2007 Acta D paper it says (p 95): "Macromolecular X-ray
crystallography is a notoriously poor method for determining the structure
of small molecules that are bound to macromolecules [...]" and then goes on
to explain why this is the case.

In the attached 2003 paper (pooling the wisdom of several of the usual
suspects, including Eleanor) it says something similar (p 1057):

"Coordinates of molecules that have been determined in complex with
macromolecules previously can of course also be retrieved from the PDB
(Bernstein et al., 1977; Berman et al., 2000), HIC-Up (Kleywegt and Jones,
1998), or CHEMPDB (Boutselakis et al., 2003). However, one should keep in
mind that these coordinates are the result of refinement against
comparatively
low-resolu- tion data where the small molecule constituted only a minute
fraction of the total scattering matter. This makes these coordinates
inherently much less accurate than those obtained from the CSD. In addition,
the coordi- nates may contain errors due to the use of incorrect restraints.

Hence, such coordinate sets should only be used as a last resort, and only
after verification that they are reliable. The latter can be facilitated by
inspection of the electron density for the compound in question, for
instance at the Uppsala Electron-Density Server (http://
fsrv1.bmc.uu.se/eds) (G.J.K. 
et al., submitted)."

Happy to be confused with George though!

--Gerard (no, the other one)



On Tue, 14 Jul 2020, Bernhard Rupp wrote:

> Hi Fellows,
>
>
>
> afaicrimps (as far as I can recall in my progressing senility)  
> someone once wrote/stated/cursed somewhere that "Macromolecular 
> refinement is not a small molecule structure determination method".
>
>
>
> Any citable source - George Sheldrick might be a suspect.
>
>
>
> Thanks & best regards, BR
>
>
>
> --
>
> Bernhard Rupp
>
>  http://www.hofkristallamt.org/
>
>  b...@hofkristallamt.org
>
> +1 925 209 7429
>
> +43 676 571 0536
>
> --
>
> Many plausible ideas vanish
>
> at the presence of thought
>
> --
>
>
>
>
> ##
> ##
>
> 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/
>


Best wishes,

--Gerard

**
Gerard J. Kleywegt

   http://xray.bmc.uu.se/gerard   mailto:ger...@xray.bmc.uu.se
**
The opinions in this message are fictional.  Any similarity
to actual opinions, living or dead, is purely coincidental.
**
Little known gastromathematical curiosity: let "z" be the
radius and "a" the thickness of a pizza. Then the volume
 of that pizza is equal to pi*z*z*a !
**



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] RA position in Crysalin technology

[Michael Fairhead]

Hello,
We have the below position available in the Frank von Delft lab at the SGC 
Oxford, applications are to be made online but I am happy yo respond ta any 
questions you may have.
Cheers
Mike

https://www.jobs.ac.uk/job/CAR123/research-assistant-in-crysalin-technology


Nuffield Department of Medicine, Structural Genomic Consortium, Old Road Campus 
Research Building, Headington, Oxford

Grade 6: £29,176 - £34,804 p.a.

Applications are invited for a Research Associate in Crysalin Technology to 
work within Protein Crystallography group at the Structural Genomic Consortium.

You will provide support to research to develop the Crysalin scaffolds into a 
routine route to obtaining crystal and cryoEM structures, by taking an 
intelligent brute-force approach.  You will work with senior scientists to 
develop and then drive a process that will bring at least 100 constructs to 
crystallisation and then diffraction every month, using this to discover the 
best combination of components and mutations to achieve the desired scaffolds.  
You will also provide advice and guidance to research students where 
appropriate.

You will hold a BSc degree or equivalent experience in molecular biology, cell 
biology, biochemistry or structural biology alongside with proven experience in 
molecular cloning and protein expression and purification. You will have the 
ability to manage own research and administrative activities. You will have the 
ability to organise and prioritise work efficiently, delivering results to a 
high standard and to an agreed schedule. You will be able to demonstrate a high 
level of communication skills, both written and oral, and the ability to work 
effectively within a team.

The position is funded by the BBSRC and is fixed-term until 31 December 2022 in 
the first instance.

Applications for this vacancy are to be made online. You will be required to 
upload a CV and supporting statement as part of your online application. Please 
quote reference 146871 on all correspondence.

Only applications received before 12.00 noon on 29 July 2020 will be considered.

https://my.corehr.com/pls/uoxrecruit/erq_jobspec_version_4.jobspec?p_id=146871

Closing Date: 29-JUL-2020 12:00


Michael Fairhead
PX Group (Frank von Delft)
SGC Oxford
Old Research Campus Building
Roosevelt Drive
Headington
OX3 7DQ
UK
Office: +44 (0) 1865 617577
Mobile: +44 (0) 7484 198808
https://www.researchgate.net/profile/Michael_Fairhead




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] Multiple Postdoctoral Fellowships available at the “Multiscale Research Institute for Complex Systems” at Fudan University of Shanghai

[lyguo]

The Multiscale Research Institute for Complex Systems (MRICS) at Fudan 
University is located at the Zhangjiang Campus of Fudan University and is 
supported by the Shanghai High-level Talents Program.  MRICS is strongly 
committed to the development of novel and effective multi-scale imaging 
technology that spans microscopic molecular structures all the way to 
macroscopic medical imaging, with the aim to provide unprecedented spatial and 
temporal insights into the structures and functions of living beings at all 
levels (molecules, cells, tissues, organs and even whole organisms).  
Specifically for structural biology, MRICS is equipped with a state-of-the-art 
cryo-EM facility that includes FEI Titan Krios with Volta phase plate, Glacios, 
Talos and Aquilos.  MRICS is also located next to Shanghai Synchrotron 
Radiation Facility for X-ray crystallography.
 
Our team includes Nobel laureate and international leading interdisciplinary 
experts.
 
We have new openings for multiple postdoctoral fellows in structural biology 
who will mainly study important biological systems by means of cryo-electron 
microscopy including single-particle and tomography. 
   
Requirements: 
 
The applicants should have a recent Ph.D. degree (within two years of 
graduation) or will have a Ph.D. degree within the next six months in biology 
or chemistry-related fields, be devoted to excellence in scientific research, 
have strong sense of responsibility, and be highly motivated and hardworking.  
For these positions, extensive experience in protein expression and 
purification is a must, while prior experience in X-ray crystallography or 
cryo-EM is a plus, but is not required.  
 
Compensation: 
 
1)We offer internationally competitive salary, fringe benefits and yearly 
bonus.  The level of salary will be determined according to the applicant's 
experience and qualification; 
2)Low-rent housing on campus is provided; 
3)There are ample opportunities to collaborate with world-renown 
laboratories;
4)We provide support for applying for funding opportunities whenever 
applicable.  
 
Shanghai is one of the most cosmopolitan cities in China with strong economy 
and vibrant scientific community. 
 
For interested applicants, please submit postdoctoral application packages (a 
combined pdf) including resumes, representative publications, phone numbers and 
email addresses of three academic referees to mrics...@fudan.edu.cn.
 
We look forward to your joining of our first-class team!



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/