[chimerax-users] hbonds and the number of detected interactions

Fri Jan 28 09:36:20 PST 2022

The "slop" options are explained in the help:
<https://rbvi.ucsf.edu/chimerax/docs/user/commands/hbonds.html#options>

The paper cited in the "hbonds" help page linked above has many different geometric criteria for different types of atoms.  

If (for example) that paper says the distance limit for a specific pair of atoms is 2.9 angstroms and the distSlop is 0.4 then it means to find ones that are up to 3.3 angstroms apart.  The orange color then means the distance was over 2.9 but within 3.3.

What slop values to use is your decision.  We made the defaults to be reasonable in most situations.

Elaine

> On Jan 28, 2022, at 1:06 AM, Enrico Martinez via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
> 
> Thank you very much, Eric!
> In fact I have already read it but there are some things not clear for me.
> May you give me a small example how I could compare the slope values
> of chimera with the geometric criteriums shown in the paper ? Does the
> dist slope correspond to the r (A) ?
> And finally how could we explain that almost all predicted h-bonds are
> colored in the second (orange) colour using default slops? Assuming
> that I am working with a structure of very good resolution (1.4A) does
> it means that I should DECREASE the both slope values for my case?
> Many thanks in advance
> Enrico
> 
> чт, 27 янв. 2022 г. в 19:52, Eric Pettersen <pett at cgl.ucsf.edu>:
>> 
>> Hi Enrico,
>>        You should really read the paper Elaine mentioned to get a feel for the range of angles and distances for various donor/acceptor types.  You also have to consider the resolution of the structure you are using and therefore the uncertainty in atomic positions (the principal reason that there are "slop" values in the calculation).  There really is no single right answer.  You have to use your judgment.
>> 
>> --Eric
>> 
>>> On Jan 27, 2022, at 10:15 AM, Enrico Martinez via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
>>> 
>>> Many thanks, Elaine!
>>> I just wonder what may be the upper limit for the both parameters to
>>> i) increase some flexibility thus allowing to consider even weaker
>>> h-bonds but ii) to be realistic ;-)
>>> 
>>> I've just recalculated again the hbond using two colors for slop, with
>>> the default params 0.4 and 20
>>> hbonds #1 & ligand reveal true log true radius 0.06 dashes 6 color
>>> turquoise distSlop 0.4 angleSlop 20 twoColors true slopColor firebrick
>>> intraRes false
>>> and generally the most h-bonds were colored in firebrick meaning that
>>> they could be less significant..
>>> Cheers,
>>> Enrico
>>> 
>>> чт, 27 янв. 2022 г. в 18:23, Elaine Meng <meng at cgl.ucsf.edu>:
>>>> 
>>>> It is beyond the scope of this forum to discuss the physics of  individual H-bonds or how to "validate" them. People have argued for years about different kinds of H-bonds, their relative strengths, etc.   We set default parameters to give what we think is the most reasonable results in most situations, based on a detailed published survey of H-bonding structures (see publication cited in help)... but then allow users to adjust the parameters at their own discretion.
>>>> 
>>>> Elaine
>>>> 
>>>>> On Jan 27, 2022, at 1:18 AM, Enrico Martinez <jmsstarlight at gmail.com> wrote:
>>>>> 
>>>>> p.s. So based on your experience and the provided screenshot, is it
>>>>> possible to validate this hydrogen bond? I believe that Glu is a very
>>>>> strong negatively charged residue which normally provides
>>>>> salt-bridges... BTW I analyzed another X-ray structure of the same
>>>>> complex and could validate there a hydrogen bond between ligand and
>>>>> the back-bone of the same Glu-166 which seems to me more bizarre (this
>>>>> GLU belongs to the betta-strand part of the structure, where in the
>>>>> visualization usually we may see an arrow => ) :-0
>>>>> 
>>>>> ср, 26 янв. 2022 г. в 20:58, Enrico Martinez <jmsstarlight at gmail.com>:
>>>>>> 
>>>>>> I gotcha!
>>>>>> So eventually I could get the hbond with the Glu-166 (as i had already
>>>>>> observed for the same complex using py at ol) using
>>>>>> the angle slop = 40 and the distance slop= 0.6
>>>>>> :-)
>>>>>> 
>>>>>> ср, 26 янв. 2022 г. в 20:47, Elaine Meng <meng at cgl.ucsf.edu>:
>>>>>>> 
>>>>>>> There is a misunderstanding: please read the hbonds help, the link I sent before.  The "slop" values are not the cutoffs; instead they are used to expand the ranges of the cutoffs.  There is no single angle cutoff  or distance cutoff.  Instead, many different cutoffs are used for different combinations of atom types depending on their elements and hybridization states.  These multiple cutoff values are listed in several tables in a publication cited in the help page.  If you cannot access that publication let me know.
>>>>>>> 
>>>>>>> Also the default "slop" values are what we recommend for most PDB files.  If you increase the values very much, you may get very poor results, "bad" (or impossible) H-bonds.
>>>>>>> Elaine
>>>>>>> 
>>>>>>> 
>>>>>>>> On Jan 26, 2022, at 11:34 AM, Enrico Martinez <jmsstarlight at gmail.com> wrote:
>>>>>>>> 
>>>>>>>> p.s. I've just increased angle to 35 and the distance to 0.7: now
>>>>>>>> there is contact with the Thr26 which is more distant, but not with
>>>>>>>> the Glu which is on the line with the NH of the ligand .. I attached
>>>>>>>> the picture :-)
>>>>>>>> 
>>>>>>>> ср, 26 янв. 2022 г. в 20:26, Enrico Martinez <jmsstarlight at gmail.com>:
>>>>>>>>> 
>>>>>>>>> Thanks a lot, Elaine!
>>>>>>>>> may you specify abit more regarding distSlop option, the values in the
>>>>>>>>> nm ? so 0,4 corresponds to 4 A, doesn't it? What the upper limit may
>>>>>>>>> be reasonable?
>>>>>>>>> Cheers,
>>>>>>>>> Enrico
>>>>>>>>> 
>>>>>>>>> ср, 26 янв. 2022 г. в 18:48, Elaine Meng <meng at cgl.ucsf.edu>:
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>>> On Jan 26, 2022, at 8:50 AM, Enrico Martinez via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
>>>>>>>>>>> 
>>>>>>>>>>> Dear ChimeraX users!
>>>>>>>>>>> I have a question regarding calculations of the hydrogen bonds and its
>>>>>>>>>>> visualisation using hbonds. Here is my command:
>>>>>>>>>>> hbonds #1 & ligand reveal true log true radius 0.06 dashes 6 color
>>>>>>>>>>> turquoise distSlop 0.4 angleSlop 20 intraRes false
>>>>>>>>>>> 
>>>>>>>>>>> So I ignore intraRes (since the ligand is also present in the #1) and
>>>>>>>>>>> calculate hydrogen bonds using default parameters.
>>>>>>>>>>> 
>>>>>>>>>>> 1) How can I modify geometric criteria for the calculations (is this
>>>>>>>>>>> distSlop / angleSlop) to increase the distance for the consideration
>>>>>>>>>>> of potential h-bonds ?  In particular as you may see on the attached
>>>>>>>>>>> image the bond between the side-chain of the Glu166 and NH of the
>>>>>>>>>>> ligand is absent (I've just checked the same structure using py*ol and
>>>>>>>>>>> it could detect this contact).
>>>>>>>>>> 
>>>>>>>>>> Yes, angleSlop makes the angle criteria looser and distSlop makes the distance criteria looser.  Larger values = more permissive, meaning more tolerant of suboptimal H-bond geometry. These parameters are explained in the hbonds help:
>>>>>>>>>> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/hbonds.html>
>>>>>>>>>> 
>>>>>>>>>>> 2) In the case of another ligand, I could visualize two h-bonds
>>>>>>>>>>> provided by two hydrogens of the side chain of the Gln189 (same
>>>>>>>>>>> residue) with two different oxygens of the ligand. Is this correct?
>>>>>>>>>> 
>>>>>>>>>> Each possible H-bond is evaluated by itself, i.e. it is not necessarily true that all of the possible H-bonds that are found could be correct at the same time. The program will correctly use the distance and angle information it is given, but (just like almost anything in life) you should also pay attention and use your own judgment.  For example, sometimes it will find two H-bonds between the same two atoms if they could both be donor or acceptor (D->A and A<-D).  However, in the situation you mention I don't see any reason to think it is not correct. There are two hydrogens and they could certainly be pointing toward two different oxygens of the ligand.
>>>>>>>>>> 
>>>>>>>>>> I hope this helps,
>>>>>>>>>> Elaine
>>>>>>>>>> -----
>>>>>>>>>> Elaine C. Meng, Ph.D.
>>>>>>>>>> UCSF Chimera(X) team
>>>>>>>>>> Department of Pharmaceutical Chemistry
>>>>>>>>>> University of California, San Francisco
>>>>>>>>>> 
>>>>>>>> <hbond_tet.png>
>>>>>>> 
>>>>> 
>>>> 
>>> 
>>> _______________________________________________
>>> ChimeraX-users mailing list
>>> ChimeraX-users at cgl.ucsf.edu
>>> Manage subscription:
>>> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
>>> 
>> 
> 
> _______________________________________________
> ChimeraX-users mailing list
> ChimeraX-users at cgl.ucsf.edu
> Manage subscription:
> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
> 