Re: [ccp4bb] Unknown electron density blob, pdb convention for partially ordered ligands

[Robbie Joosten]

Hi Tristan,



There are PDB entries that have this, but this makes matters a bit more 
complicated in annotation. You have to define LINKs and leaving atoms. 
Consistency would be nice here, but that is lacking in these entries.



Cheers,

Robbie



Sent from my Windows 10 phone



Van: Tristan Croll<mailto:ti...@cam.ac.uk>
Verzonden: woensdag 25 januari 2017 18:30
Aan: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK>
Onderwerp: Re: [ccp4bb] Unknown electron density blob, pdb convention for 
partially ordered ligands



I've often wondered about PEG (and, I guess, other synthetic polymers): 
wouldn't it just be better to define the monomer, and then model a chain of 
however many monomers you need?

T



Tristan Croll
Research Fellow
Cambridge Institute for Medical Research
University of Cambridge CB2 0XY




> On 25 Jan 2017, at 17:21, Edward A. Berry <ber...@upstate.edu> wrote:
>
> Uma's use of quotes around "di" suggests a related question about PDB 
> convention. It was my (perhaps not very good) understanding that ligands 
> should be identified by what is actually present in the crystal, and not by 
> what can be modeled. For example endogenous ubiquinone is likely to be UQ50 
> (depending on the species) but most of that 50-carbon side chain is hanging 
> out in the lipid or detergent and completely disordered. Still we should use 
> the ligand identifier for UQ50, even though codes exist for UQ with 5 or 
> 10-carbon side chains that are much better accommodated by the density.
>
> If that is the case, one should not use the pdb identifier for diethylene 
> glycol (PEG) when PEG4k was the precipitant, unless you believe that the 
> binding site has specifically selected diethylene glycol from an extremely 
> broad range of polymer lengths in the added material.  Using the identifier 
> for a much longer PEG will result in a large number of "missing atoms" listed 
> in the report, but would eliminate the unreasonable assumption that PEG 
> fragment models must always end with a terminal oxygen.
>
> Even if that is the rule, I would agree that PEGs would be a good place to 
> ignore the rule. Since PEGs have a MW distribution, it is impossible to know 
> exactly what is bound and it may be different in different unit cells. If you 
> are not going to get it right no matter what you put, you might as well put 
> something that fits.
> eab
>
>> On 01/25/2017 09:51 AM, Uma Gabale wrote:
>> Dear all,
>> Thank you very much for your replies. It is a PEG, a "di"ethylene glycol to 
>> be precise, in most chains.
>> Best regards,
>> Uma.
>> --
>> Uma Gabale, PhD
>> Research Associate
>> Molecular and Cellular Biochemistry
>> Indiana University Bloomington
>>

Re: [ccp4bb] Unknown electron density blob, pdb convention for partially ordered ligands

[Tristan Croll]

I've often wondered about PEG (and, I guess, other synthetic polymers): 
wouldn't it just be better to define the monomer, and then model a chain of 
however many monomers you need?

T

 
 
Tristan Croll
Research Fellow
Cambridge Institute for Medical Research
University of Cambridge CB2 0XY
 

 

> On 25 Jan 2017, at 17:21, Edward A. Berry  wrote:
> 
> Uma's use of quotes around "di" suggests a related question about PDB 
> convention. It was my (perhaps not very good) understanding that ligands 
> should be identified by what is actually present in the crystal, and not by 
> what can be modeled. For example endogenous ubiquinone is likely to be UQ50 
> (depending on the species) but most of that 50-carbon side chain is hanging 
> out in the lipid or detergent and completely disordered. Still we should use 
> the ligand identifier for UQ50, even though codes exist for UQ with 5 or 
> 10-carbon side chains that are much better accommodated by the density.
> 
> If that is the case, one should not use the pdb identifier for diethylene 
> glycol (PEG) when PEG4k was the precipitant, unless you believe that the 
> binding site has specifically selected diethylene glycol from an extremely 
> broad range of polymer lengths in the added material.  Using the identifier 
> for a much longer PEG will result in a large number of "missing atoms" listed 
> in the report, but would eliminate the unreasonable assumption that PEG 
> fragment models must always end with a terminal oxygen.
> 
> Even if that is the rule, I would agree that PEGs would be a good place to 
> ignore the rule. Since PEGs have a MW distribution, it is impossible to know 
> exactly what is bound and it may be different in different unit cells. If you 
> are not going to get it right no matter what you put, you might as well put 
> something that fits.
> eab
> 
>> On 01/25/2017 09:51 AM, Uma Gabale wrote:
>> Dear all,
>> Thank you very much for your replies. It is a PEG, a "di"ethylene glycol to 
>> be precise, in most chains.
>> Best regards,
>> Uma.
>> --
>> Uma Gabale, PhD
>> Research Associate
>> Molecular and Cellular Biochemistry
>> Indiana University Bloomington
>> 

Re: [ccp4bb] Unknown electron density blob, pdb convention for partially ordered ligands

[Edward A. Berry]

Uma's use of quotes around "di" suggests a related question about PDB 
convention. It was my (perhaps not very good) understanding that ligands should be 
identified by what is actually present in the crystal, and not by what can be modeled. 
For example endogenous ubiquinone is likely to be UQ50 (depending on the species) but 
most of that 50-carbon side chain is hanging out in the lipid or detergent and completely 
disordered. Still we should use the ligand identifier for UQ50, even though codes exist 
for UQ with 5 or 10-carbon side chains that are much better accommodated by the density.

If that is the case, one should not use the pdb identifier for diethylene glycol (PEG) 
when PEG4k was the precipitant, unless you believe that the binding site has specifically 
selected diethylene glycol from an extremely broad range of polymer lengths in the added 
material.  Using the identifier for a much longer PEG will result in a large number of 
"missing atoms" listed in the report, but would eliminate the unreasonable 
assumption that PEG fragment models must always end with a terminal oxygen.

Even if that is the rule, I would agree that PEGs would be a good place to 
ignore the rule. Since PEGs have a MW distribution, it is impossible to know 
exactly what is bound and it may be different in different unit cells. If you 
are not going to get it right no matter what you put, you might as well put 
something that fits.
eab

On 01/25/2017 09:51 AM, Uma Gabale wrote:

Dear all,
Thank you very much for your replies. It is a PEG, a "di"ethylene glycol to be 
precise, in most chains.
Best regards,
Uma.
--
Uma Gabale, PhD
Research Associate
Molecular and Cellular Biochemistry
Indiana University Bloomington