[ccp4bb] resolution limits
Dear crystallographers, I have two datasets that were merged/scaled using ccp4's aimless, with resolution ranges of 52-1.7 and 57-1.9. However, upon refinement, the resolution range used by phenix.refine is 36-1.7 for one and 104-1.9 for the other. 1) Why does phenix.refine change the low resolution limits of my processed data? 2) How can I prevent this? Thanks, Andrew
[ccp4bb] waters with positive FoFc peaks?
Hello all, I have a 1.8A structure (Rfree/Rwork = 20.5/17.4) with 1420 water molecules modelled. There are approximately a dozen waters, all well structured with hydrogen bonds to protein atoms, with positive difference density (>4sigma, sometimes >5) at their centre. The only ions present in my buffer were Cs, Cl and a small amount of Na. I thought Cl ions might be a possibility, but many of them are in close proximity to acidic residues and/or one-another. It's probably worth noting that the same structure solved using data to 2.25A from the same crystal at a different wavelength doesn't contain these peaks (the offending waters look normal). Has any come across this before? Thoughts? Thanks, Andrew Marshall PhD Candidate Laboratory of Protein Crystallography Dept. of Molecular and Cellular Biology School of Biological Sciences The University of Adelaide
Re: [ccp4bb] Atom clashes in active site?
Hi Scott, That would be great if you have some references handy? Thanks very much, Andrew Marshall PhD Candidate Laboratory of Protein Crystallography Dept. of Molecular and Cellular Biology School of Biological Sciences The University of Adelaide On Wed, Dec 21, 2016 at 1:48 AM, Scott Horowitz <horow...@umich.edu> wrote: > Hi Andrew, > > Based on the atoms and distances you are mentioning, these don't sound > like steric clashes, but like a chalcogen bond between the S and O atoms, > and CH...O hydrogen bonds between the O and CH3. These are common and > well-accepted interactions, but unfortunately aren't usually treated as > such by refinement programs. Let me know if you want references for these > interaction types. > > Scott > > On Mon, Dec 19, 2016 at 8:48 PM, Andrew Marshall < > andrew.c.marsh...@adelaide.edu.au> wrote: > >> Hi all, >> >> Thank you for your suggestions. I tried the pdb file edit (making the >> offending atoms of both the ligand and the protein 'B' altconf), but it >> didn't seem to make any difference to their positions after a single round >> of refinement..? >> The atoms in the active site concern two acetyl groups - one from the >> substrate, acetyl-CoA, and the other from an acetylated cysteine in the >> protein - that I believe are poised ready for a condensation reaction. The >> closest contacts are between S and O(carbonyl) atoms (2.9A) and O and CH3 >> (3.1A), but going off the density, I think these should be closer (more >> like 2.8 or 2.7A). It may be that I've trapped another reaction >> intermediate (which would be cool), but I don't think that fits the density >> quite as well. Any thoughts/ideas? >> >> Thanks, >> >> Andrew Marshall >> PhD Candidate >> Laboratory of Protein Crystallography >> Dept. of Molecular and Cellular Biology >> School of Biological Sciences >> The University of Adelaide >> >> On Tue, Dec 20, 2016 at 12:09 AM, Scott Horowitz <horow...@umich.edu> >> wrote: >> >>> Hi Andrew, >>> >>> I'm curious- what are the atoms that are clashing? I worked on this sort >>> of thing back in my Ph.D., and so I might have some useful tidbits if, for >>> example, the S is clashing with a carbon of some sort. >>> >>> Thanks, >>> Scott >>> >>> On Mon, Dec 19, 2016 at 12:39 AM, Andrew Marshall < >>> andrew.c.marsh...@adelaide.edu.au> wrote: >>> >>>> Hi all, >>>> >>>> I have a structure of a condensing enzyme with substrate bound. The >>>> active site is very tight, requiring some of the substrate atoms to clash >>>> with a catalytic cysteine. This means that although the substrate fits the >>>> density nicely upon manual real-space refinement, phenix recognises the >>>> clash, resulting in the displacement of substrate atoms so that they are >>>> outside the density. I can mostly fix this by using distance restraints, >>>> but I'd rather allow it to refine in a less biased manner, but ignore the >>>> clash. Is this a acceptable way forward? If so, is there a parameter I can >>>> edit to tell phenix to ignore clashes between these specific atoms? >>>> >>>> Thanks, >>>> >>>> Andrew Marshall >>>> PhD Candidate >>>> Laboratory of Protein Crystallography >>>> Dept. of Molecular and Cellular Biology >>>> School of Biological Sciences >>>> The University of Adelaide >>>> >>>> >>> >>> >>> -- >>> Scott Horowitz, Ph.D. >>> Postdoctoral Fellow >>> >>> University of Michigan >>> Department of Molecular, Cellular, and Developmental Biology >>> Bardwell lab >>> 830 N. University Ave, Room 4007 >>> Ann Arbor, MI 48109 >>> phone: 734-647-6683 >>> fax: 734-615-4226 >>> >> >> > > > -- > Scott Horowitz, Ph.D. > Postdoctoral Fellow > > University of Michigan > Department of Molecular, Cellular, and Developmental Biology > Bardwell lab > 830 N. University Ave, Room 4007 > Ann Arbor, MI 48109 > phone: 734-647-6683 > fax: 734-615-4226 >
Re: [ccp4bb] Atom clashes in active site?
Hi Pavel,
That worked a treat! Thanks again for your help,
Andrew Marshall
PhD Candidate
Laboratory of Protein Crystallography
Dept. of Molecular and Cellular Biology
School of Biological Sciences
The University of Adelaide
On Tue, Dec 20, 2016 at 3:18 PM, Pavel Afonine <pafon...@gmail.com> wrote:
> Hi Andrew,
>
> yes, exactly as you describe: set distance_ideal to a meaningful value
> (approx. distance between density peaks, doesn't have to be very accurate),
> and set sigma to some large number, say 1 or so.
>
> Pavel
>
>
> On Mon, Dec 19, 2016 at 7:40 PM, Andrew Marshall <
> andrew.c.marsh...@adelaide.edu.au> wrote:
>
>> Hi Pavel,
>>
>> To define a weak bond, would you use "geometry_restraints.edits { ...
>> bond {... " , and just set a rough distance_ideal and a very high sigma
>> (like say 5A)?
>> Or are you referring to something different?
>>
>> Andrew Marshall
>> PhD Candidate
>> Laboratory of Protein Crystallography
>> Dept. of Molecular and Cellular Biology
>> School of Biological Sciences
>> The University of Adelaide
>>
>> On Tue, Dec 20, 2016 at 1:28 PM, Pavel Afonine <pafon...@gmail.com>
>> wrote:
>>
>>> Hi Andrew,
>>>
>>> you can define a weak bond between clashing atoms which will disable
>>> repulsion. A weak bond should not introduce any bias.
>>>
>>> Pavel
>>>
>>> On Sun, Dec 18, 2016 at 9:39 PM, Andrew Marshall <
>>> andrew.c.marsh...@adelaide.edu.au> wrote:
>>>
>>>> Hi all,
>>>>
>>>> I have a structure of a condensing enzyme with substrate bound. The
>>>> active site is very tight, requiring some of the substrate atoms to clash
>>>> with a catalytic cysteine. This means that although the substrate fits the
>>>> density nicely upon manual real-space refinement, phenix recognises the
>>>> clash, resulting in the displacement of substrate atoms so that they are
>>>> outside the density. I can mostly fix this by using distance restraints,
>>>> but I'd rather allow it to refine in a less biased manner, but ignore the
>>>> clash. Is this a acceptable way forward? If so, is there a parameter I can
>>>> edit to tell phenix to ignore clashes between these specific atoms?
>>>>
>>>> Thanks,
>>>>
>>>> Andrew Marshall
>>>> PhD Candidate
>>>> Laboratory of Protein Crystallography
>>>> Dept. of Molecular and Cellular Biology
>>>> School of Biological Sciences
>>>> The University of Adelaide
>>>>
>>>>
>>>
>>
>
Re: [ccp4bb] Atom clashes in active site?
Hi Pavel,
To define a weak bond, would you use "geometry_restraints.edits { ... bond
{... " , and just set a rough distance_ideal and a very high sigma (like
say 5A)?
Or are you referring to something different?
Andrew Marshall
PhD Candidate
Laboratory of Protein Crystallography
Dept. of Molecular and Cellular Biology
School of Biological Sciences
The University of Adelaide
On Tue, Dec 20, 2016 at 1:28 PM, Pavel Afonine <pafon...@gmail.com> wrote:
> Hi Andrew,
>
> you can define a weak bond between clashing atoms which will disable
> repulsion. A weak bond should not introduce any bias.
>
> Pavel
>
> On Sun, Dec 18, 2016 at 9:39 PM, Andrew Marshall <
> andrew.c.marsh...@adelaide.edu.au> wrote:
>
>> Hi all,
>>
>> I have a structure of a condensing enzyme with substrate bound. The
>> active site is very tight, requiring some of the substrate atoms to clash
>> with a catalytic cysteine. This means that although the substrate fits the
>> density nicely upon manual real-space refinement, phenix recognises the
>> clash, resulting in the displacement of substrate atoms so that they are
>> outside the density. I can mostly fix this by using distance restraints,
>> but I'd rather allow it to refine in a less biased manner, but ignore the
>> clash. Is this a acceptable way forward? If so, is there a parameter I can
>> edit to tell phenix to ignore clashes between these specific atoms?
>>
>> Thanks,
>>
>> Andrew Marshall
>> PhD Candidate
>> Laboratory of Protein Crystallography
>> Dept. of Molecular and Cellular Biology
>> School of Biological Sciences
>> The University of Adelaide
>>
>>
>
Re: [ccp4bb] Atom clashes in active site?
Hi all, Thank you for your suggestions. I tried the pdb file edit (making the offending atoms of both the ligand and the protein 'B' altconf), but it didn't seem to make any difference to their positions after a single round of refinement..? The atoms in the active site concern two acetyl groups - one from the substrate, acetyl-CoA, and the other from an acetylated cysteine in the protein - that I believe are poised ready for a condensation reaction. The closest contacts are between S and O(carbonyl) atoms (2.9A) and O and CH3 (3.1A), but going off the density, I think these should be closer (more like 2.8 or 2.7A). It may be that I've trapped another reaction intermediate (which would be cool), but I don't think that fits the density quite as well. Any thoughts/ideas? Thanks, Andrew Marshall PhD Candidate Laboratory of Protein Crystallography Dept. of Molecular and Cellular Biology School of Biological Sciences The University of Adelaide On Tue, Dec 20, 2016 at 12:09 AM, Scott Horowitz <horow...@umich.edu> wrote: > Hi Andrew, > > I'm curious- what are the atoms that are clashing? I worked on this sort > of thing back in my Ph.D., and so I might have some useful tidbits if, for > example, the S is clashing with a carbon of some sort. > > Thanks, > Scott > > On Mon, Dec 19, 2016 at 12:39 AM, Andrew Marshall < > andrew.c.marsh...@adelaide.edu.au> wrote: > >> Hi all, >> >> I have a structure of a condensing enzyme with substrate bound. The >> active site is very tight, requiring some of the substrate atoms to clash >> with a catalytic cysteine. This means that although the substrate fits the >> density nicely upon manual real-space refinement, phenix recognises the >> clash, resulting in the displacement of substrate atoms so that they are >> outside the density. I can mostly fix this by using distance restraints, >> but I'd rather allow it to refine in a less biased manner, but ignore the >> clash. Is this a acceptable way forward? If so, is there a parameter I can >> edit to tell phenix to ignore clashes between these specific atoms? >> >> Thanks, >> >> Andrew Marshall >> PhD Candidate >> Laboratory of Protein Crystallography >> Dept. of Molecular and Cellular Biology >> School of Biological Sciences >> The University of Adelaide >> >> > > > -- > Scott Horowitz, Ph.D. > Postdoctoral Fellow > > University of Michigan > Department of Molecular, Cellular, and Developmental Biology > Bardwell lab > 830 N. University Ave, Room 4007 > Ann Arbor, MI 48109 > phone: 734-647-6683 > fax: 734-615-4226 >
[ccp4bb] Atom clashes in active site?
Hi all, I have a structure of a condensing enzyme with substrate bound. The active site is very tight, requiring some of the substrate atoms to clash with a catalytic cysteine. This means that although the substrate fits the density nicely upon manual real-space refinement, phenix recognises the clash, resulting in the displacement of substrate atoms so that they are outside the density. I can mostly fix this by using distance restraints, but I'd rather allow it to refine in a less biased manner, but ignore the clash. Is this a acceptable way forward? If so, is there a parameter I can edit to tell phenix to ignore clashes between these specific atoms? Thanks, Andrew Marshall PhD Candidate Laboratory of Protein Crystallography Dept. of Molecular and Cellular Biology School of Biological Sciences The University of Adelaide
