# AMBER Archive (2005)Subject: Re: AMBER: Question about PDB format

From: Andy Purkiss (a.purkiss_at_mail.cryst.bbk.ac.uk)
Date: Sat Nov 19 2005 - 13:19:00 CST

The pdb format is based on the old fortran card format 8F3 starting with the
x-coordinate in column 31 and then the y in 39 and the z in 47. This means that
the length of each coordinate entry is 8 characters with 3 digits after the
decimal place. So the maximum number allowed is 9999.999 and the minimum
-999.999.

As leap is C-code, it does not rely on the fixed width columns of fortran-77 and
so the pdb format can be twisted by shifting the decimal point to the left one
position to give 8f4 numbers, max 999.9999, minimum -99.9999. So you might get
into trouble with negative coordinates using the strategy you have below. Note
that the coordinates might have no spaces between them and this can cause
problems with some code with doesn't use the column numbers.

This might be easier to see with the following column number list (FIXED FORMAT
NEEDED!)

Coordinate <- x -><- y -><- z ->
Column 333333333444444444455555
Number 123456789012345678901234

Max 8f3 9999.9999999.9999999.999
Min 8f3 -999.999-999.999-999.999
Max 8f4 999.9999999.9999999.9999
Min 8f4 -99.9999-99.9999-99.9999
3Decimal 0.000 0.000 0.000 1.00 0.00
1 1.000 0.000 0.000 1.00 0.00
1 -0.334 0.943 0.000 1.00 0.00
'problem 0.0001 0.0001 0.0001 1.00 0.00
4Decimal' 1.0001 0.0001 0.0001 1.00 0.00
1 -0.3341 0.9431 0.0001 1.00 0.00
4Decimal 0.0001 0.0001 0.0001 1.00 0.00
1 1.0001 0.0001 0.0001 1.00 0.00
1 -0.3341 0.9431 0.0001 1.00 0.00

Note how the additional 1s for the x-coordinate are in the start of the
y-coordinate entry.

The way around this sort of problem is to upgrade our software to use the XML or
mmCIF formats that the pdb provides. The crystallographers are moving (slowly)
in this direction. XML and/or mmCIF do not have the same fixed format limits of
the old fortran card system and allow much more data to be stored in the
files.

Hope this is of some help.

Andy Purkiss

Quoting FyD <fyd_at_u-picardie.fr>:

> Dear All,
>
> I have a question about the PDB format recognized by LEaP...
>
> The PDB format is described @
> http://www.rcsb.org/pdb/docs/format/pdbguide2.2/guide2.2_frame.html
> And the Cart. coord. follow the following format:
> 1 2 3 4 5 6 7
> 8
>
12345678901234567890123456789012345678901234567890123456789012345678901234567890
> MODEL 1
> ATOM 1 N ALA 1 11.104 6.134 -6.504 1.00 0.00
> N
> ATOM 2 CA ALA 1 11.639 6.071 -5.147 1.00 0.00
> C
>
> If I generate a PDB structure using the builder in LEaP (H2O), I get:
> ATOM 1 O1 HHO 1 0.000 0.000 0.000 1.00 0.00
> ATOM 2 H2 HHO 1 1.000 0.000 0.000 1.00 0.00
> ATOM 3 H3 HHO 1 -0.334 0.943 0.000 1.00 0.00
>
> Now, I decide to add a fourth digit to this water molecule:
> ATOM 1 O1 HHO 1 0.0001 0.0001 0.0001 1.00 0.00
> ATOM 2 H2 HHO 1 1.0001 0.0001 0.0001 1.00 0.00
> ATOM 3 H3 HHO 1 -0.3341 0.9431 0.0001 1.00 0.00
> If, I try to reload it in xLEaP. The format is not anymore recognized (water
> becomes linear)...
>
> Now, I shift all the Cart. coord. one character to the left:
> ATOM 1 O1 HHO 1 0.0001 0.0001 0.0001 1.00 0.00
> ATOM 2 H2 HHO 1 1.0001 0.0001 0.0001 1.00 0.00
> ATOM 3 H3 HHO 1 -0.3341 0.9431 0.0001 1.00 0.00
> If, I try to reload this last struct in xLEaP, the format is this time
> recognized.
>
> I think there is the same 'problem' in VMD, but not in molekel...
>
> Why should I shift all the Cart. coord. to the left to get something that
> makes
> sense when I want to use four digits in LEaP ? For me, this does not follow
> the
> official PDB format anymore (I wonder if this problem was not present in
> Amber7)...
>
> Thank you, regards, Francois
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

```--
Your refrigerators and sofas will be utilized for our comfort.
Your society will be assimilated to nurture and care for our own.
Resistance is Furry."
+---------------------------------------------------------------------+
|  Andy Purkiss, School of Crystallography, Birkbeck College, London  |
|           E-mail   a.purkiss_at_mail.cryst.bbk.ac.uk                   |
|      Phone 020 7631 6869 (Work) or 020 7237 1165 (Home)             |
|                     Mobile: 07763 490 360                           |
+---------------------------------------------------------------------+
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
```