[Jmol-developers] Re: Jmol improvements in PDB element type recognition

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

>> However, this does not explain the problem with 1EBL. This entry had the
>> element type in the appropriate columns. Therefore, the
>> 'ambiguous-atom-name' problem was not the cause.
>
> I think "yes", it is the cause.  Up to now we considered only coloumn 14
> for the element difinition.

Hmmm ... not within Jmol.

The PDB file format specification for ATOM says:

13 - 16  Atom        name      Atom name.

77 - 78  LString(2)  element   Element symbol, right-justified.

Jmol has always tried to use columns 77 & 78 to determine the element
type. If they are not present, then it uses columns 13 & 14.

I pulled 1EBL from the pdb and that entry includes the element symbol in
columns 77 & 78. I ran a version of Jmol from a few days ago and did not
see a problem with element recognition.

So, it seems to me that it is still unexplained.

Q: Do we know exactly what problem David was seeing?

> Howevere PDB uses coloumn 13 for defining metal atoms.
> For example Fe has F in 13 and E in 14; Ca has C in 13 and
> A in 14 and (IMPORTANT!) Mn has M in 13 and N in 14 (as nytrogen!),
> however program should recognize Mn as Mn but not as N...

With the use of the FORMUL records to determine which elements are present
I believe that the difficulties will occur much less frequently.

Note that this FORMUL record mechanism is only used when the element type
is not present in cols 77 & 78.


Miguel