[Jmol-users] Re: CONECT records in Jmol

[Miguel <mi...@jm...>]

Regarding CONECT records in PDB files ...

CONECT records can be used to specify hydrogen bonds.

In addition, they can be used to specify double-bonds, triple-bonds, and
bonds that would not normally be identified by the auto-bonding algorithm=

because of distance.

>>>>You can define the hbonds in the .pdb file in the CONECT records.
>>>>
> <snip>
>>
>>http://www.rcsb.org/pdb/docs/format/pdbguide2.2/guide2.2_frame.html
>>
> <snip>
>>The pdb specification requires that the bonds be
>>'double-listed', in that an hbond from 1--99 must
>>also be listed as 99--1
>>
>>(in the notes: * The connectivity list given here is redundant in
>>that each bond indicated is given twice, once with each of the two
>>atoms involved specified in columns 7 - 11.)
>>
>>However, Jmol does not require/enforce this ... you only have to
>>list the hbonds once and Jmol will do the right thing.
>>
>>The other useful thing for CONECT records in .pdb files is that you
>>can use them to define double and triple bonds. Just list the
>>destination atom more than once in the 4 fields from columns 12-31
>>and Jmol will do the right thing.
>>
>
> IIRC Chime handling of CONECT records was a pain, which is why I always=

> avoided them.  you had to have *all* of the bonds in the structure as
> CONECT records, or else Chime would ignore all of them.  or something l=
ike
> that.

This is an area where Jmol is not Chime-compatible ...

I think Jmol does the right thing ;-)

> how does Jmol behave when it sees CONECT records?
> does it run its built-in
> auto-bonding routine no matter what?

There is special treatment for pdb files in order to comply with the PDB
specification ... where CONECT records generally are used only used for
unusual/special bonds, and autobonding takes care of the 'normal' bonds.

Jmol *always* respects connectivity information in underlying data files,=

regardless of file type. Autobonding (if any) comes second.

In the case of PDB files this connectivity information is contained in
CONECT records. PDB files are special in that the file specification says=

that they contain only partial bonding information for things that are no=
t
normal. (this is clearly not the exact wording)

But, some programs export so-called PDB files with complete bonding
information specified in CONECT records. So, Jmol tries to make the right=

decision about whether or not complete connectivity information was
contained in the .pdb file.

The sequence for .pdb files is as follows:

 - Process all bonds specified by CONECT records.
 - if the number of bonds is less than half the number
   of atoms (bondCount < atomCount/2) then run the
   auto-bonding to *supplement* any bonding that was
   specified in CONECT records

> and what happens if you
> have a single CONECT record for a bond that Jmol has already
> auto-generated?  does it become now a double bond?

    * if, during the auto-bonding process, a bond already
      exists between the two atoms, then the preexisting
      bond is retained. So, the autobond process will not
      turn a CONECT double-bond into a triple-bond.

 - else it was the case that (bondCount >=3D atomCount/2),
   so assume that the CONECT records fully specify all
   connectivity and do not run autobonding at all.

This means that:

if you have no CONECT records, then all the connectivity will be
determined by autobonding

if you have some CONECT records, then the connectivity will be a mixture
of the CONECT records and the autobonding, where the CONECT records take
precedence

if you have many CONECT records (more-or-less as many CONECT records as
ATOMs), then the connectivity is solely determined by the CONECT records
and autobonding is not run.


I think Jmol does the right thing ;-)


Miguel

-----
Open Source Molecular Visualization
www.jmol.org
miguel=40jmol.org
-----