Jan,
I appreciate your taking the time to write up a detailed response.
Since I am not a chemist, I had a hard time understanding some of your
answers/examples/recommendations. But we will work through them.
I have more comments/question below, but right now, the important thing
for me to understand is:
THE BIG QUESTION
----------------
What do I do when I see a .pdb file that has 200 atoms and 20 CONECT records?
> the RasMol 2.7.2.1 online help says:
>
> Connect
> Syntax: connect {<boolean>}
>
> The RasMol 'connect' command is used to force RasMol to (re)calculate
> the connectivity of the current molecule. If the original input file
> contained connectivity information, this is discarded. The command
> 'connect false' uses a fast heuristic algorithm that is suitable for
> determining bonding in large bio-molecules such as proteins and nucleic
> acids. The command 'connect true' uses a slower more accurate algorithm
> based upon covalent radii that is more suitable to small molecules
> containing inorganic elements or strained rings. If no parameters are
> given, RasMol determines which algorithm to use based on the number of
> atoms in the input file. Greater than 255 atoms causes RasMol to use the
> faster implementation. This is the method used to determine bonding, if
> necessary, when a molecule is first read in using the 'load' command.
>
> on load, RasMol uses the CONECT record stored in the PDB file, if there
> are none, RasMol calculates the connection information.
Jan,
My concern is that this seems to conflict with the PDB spec. My reading of
the PDB spec says that you only need to have CONECT records for the
'extra' bonds.
I don't have any examples in front of me but ...
Some .pdb files (like those generated from other formats), seem to have
all the connectivity defined by CONECT records.
But I have seen other .pdb files with hundreds of atoms but only a few
dozen CONECT records ... in accordance with the PDB spec.
> make your own test with 1ndh
I will check this one out
> load pdb pdb1ndh.ent
> restrict fad
Q: What is 'fad' ? some type of residue?
> # the connections look well
> # and this is because of the CONECT records
> # as you could demonstrate by:
>
> save pdb fad.pdb
(implementing 'save' is not high on the Jmol priority list)
> # look into the file, it lacks CONECT records
> # this should be done better with Jmol :-)
>
> load pdb fad.pdb
>
> # the adenine ring looks really awful
> # but this is a completely other RasMol bug
> # and Jmol seams to suffer the same bug
> # have you recognized that all ring atoms are pink
In Jmol, if the atoms are pink it is usually because the element symbol is
not recognized. These pink atoms also have a large covalent bonding radius
... so they tend to bond with lots of things ... so they are easy to see
:-)
> # there is a exception of the PDB file format with ?nearly? all these #
> nucleotide derivate ligands as FAD NAD ... which should be
> # http://www.rcsb.org/pdb/docs/format/pdbguide2.2/part_67.html
> # COLUMNS 13 - 16 Atom name Atom name. #
> where the element symbol should be right justified at position 14 and #
> http://www.rcsb.org/pdb/docs/format/pdbguide2.2/part_62.html
> # <<###
> * Columns 77 - 78 contain the atom's element symbol (as given in the
> periodic table), right-justified.
Unfortunately, most of the .pdb files that I have seen do *not* follow
this format. They are still in the older format, where these columns
contain parts of the residue ID.
I asked on the .pdb mailing list about a way to detect pdb v2 from pdb v1.
I received some complicated answers, but no simple answers.
I'll take another look at it.
> This is especially needed because in
> some cases it has not been possible to follow the convention that
> columns 13 - 14 of the atom name contain the element symbol. The most
> common cases are:
>
> - In large het groups it sometimes is not possible to follow the
> convention of having the first two characters be the chemical
> symbol and still use atom names that are meaningful to users. A
> example is nicotinamide adenine dinucleotide, atom names begin with
> an A or N, depending on which portion of the molecule they appear
> in, e.g., AC6 or NC6, AN1 or NN1.
>
> - Hydrogen naming sometimes conflicts with IUPAC conventions. For
> example, a hydrogen named HG11 in columns 13 - 16 is differentiated
> from a mercury atom by the element symbol in columns 77 - 78.
> Columns 13 - 16 present a unique name for each atom.
Sorry, I am getting lost.
> ###
> # but there is no element symbol in 1ndh
> # therefore the atom radii of Ac and An are used
>
> cpk
> cpk off
>
> # and using "exact calculation" results in garbage while
>
> connect false
My understanding of the RasMol doc is that connect true/false si only used
for performance reasons. They make this distinction between small
molecules (<255) and large molecules.
Jmol does not do this. There is one algorithm that always runs, regardless
of size.
>
> # the heuristic, gives good results
> # as is achieved if you edit 1ndh.pdb
> # s/ A(...) FAD 273/ \1 FAD 273/g
> # then the element type is recognized correctly
> # and on loading (connect true) gives good results
Sorry, I don't understand this. We will have to take it one small step at
a time.
> Jmol should deal with these ligands as they are very common
> and it should be possible to reset Jmol to the PDB files CONECT record
> information, e.g. set connect file
Reverting to the connectivity specified in the file would be easy to do
and it sounds like something we should support.
>> Q: How do I tell if the CONECT records are for 'all' the bonds or only
>> for the 'nonstandard' bonds?
>>
> use CONECT if present and calculate the rest
I'm sorry, I don't understand your response. By 'calculate the rest' to
you mean 'calculate the rest of the bonds in that molecule' or 'calculate
the rest of the files ... those that do not have CONECT records'
If there are no CONECT records in the file then it is easy ... Jmol does
that today.
If I see a file with 200 atoms and 400 CONECT records, that is probably
pretty easy too.
THE BIG QUESTION
----------------
What do I do when I see a .pdb file that has 200 atoms and 20 CONECT records?
>> Q: It seems I recall seeing something about a format 'extension' where
>> a CONECT record could be specified more than once in order to define a
>> double/triple bond. Is that true?
>>
> don't know, I used
> bond <number> <number> +
> on an existing bond (or twice) to alter the wireframe representation
>
>
>>
>> Another factor ... the RasMol documentation for the 'connect' command
>> says:
>> The RasMol 'connect' command is used to force RasMol to
>> (re)calculate the connectivity of the current molecule.
>> If the original input file contained connectivity
>> information, this is discarded.
>>
>> So, that doesn't seem right to me in the case where a file is
>> following the PDB format specification and standard connectivity is
>> not included in the CONECT records.
>>
>> Q: What should the connect command do?
>>
> connect [false|off|1.54]
> calculate connections according to heuristic, e.g. assuming to deal with
> average C N O bond length which may be less then e.g. 1.98 ?who knows? I
> don't know if it is simply bond length which is used by RasMol
> heuristic, but if, I'll suggest to allow, in addition to the parameters
> false or off, which RasMol uses, to set the distance e.g. 1.54 which is
> said to be the average C-C or 1.10 for N%N (nitrogen:-)
Sorry, I didn't follow this.
> connect [true|on]
> to use all information about an atom e.g. atom radius to calculate
> bonding
I *think* I understand this one. Use covalent radii of the atoms to
calculate bonding.
> connect file
> which resets to the state after file loading (uses CONECT records)
What do I do if there were no CONECT records in the file?
Miguel
>
> Regards, Jan
>
>>
>
> --
> Snail to:Jan Reichert, IMB, D-07708 Jena, POB 100813.
> Voice:+49 3641 6562-05 (Fax:+49 3641 6562-10) :/i
> mailto:[EMAIL PROTECTED]
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive? Does it
> help you create better code? SHARE THE LOVE, and help us help
> YOU! Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Jmol-users mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/jmol-users
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/
_______________________________________________
Jmol-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jmol-users
