PDB hybrid36 format (Packmol PDB files with hexadecimal numbers)

[richardjgowers]

Here's a really funny one packmol writes PDB indices in hexadecimal when they get too high. PDB really is the format that never stops giving

[kain88-de]

https://www-s.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2015-2016/0569.html

VMD might do something similar.

[kain88-de]

http://thegrantlab.org/bio3d/html/read.pdb.html yup now I'm pretty sure VMD supports hex numbers. VMD will also be an easy way to generate a test file

[richardjgowers]

Ok but packmol only numbers hexidecimally *after* 99,999 😂 maybe that’s a bug to report

…

On Sat, May 12, 2018 at 11:00 AM, Max Linke ***@***.***> wrote: http://thegrantlab.org/bio3d/html/read.pdb.html yup now I'm pretty sure VMD supports hex numbers. VMD will also be an easy way to generate a test file — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#1897 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AI0jB4wcLv2ZWqUZtK2PT5dph7WsfdmLks5txvkngaJpZM4T8cMh> .

[kain88-de]

I'm not sure if VMD doesn't do that too.

[richardjgowers]

Ah apparently VMD does the same. Which means we'd have to read more than 16 lines to check this. I think this will be fairly easy to fix, but harder to fix with no performance loss, which is preferable as huge PDB files are commonly used.

[richardjgowers]

Yeah from what I read that’s the standard... odd.

…

On Sat, May 12, 2018 at 11:56 AM, Max Linke ***@***.***> wrote: I'm not sure if VMD doesn't do that too. — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#1897 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AI0jB1JieA-it6A4iFIf8wPveKV_6-Wxks5txwYggaJpZM4T8cMh> .

[orbeckst]

There's a version of the "PDB format" that uses hexadecimals called hybrid-36. (This was discussed in openmm/openmm#659 a while ago.) We can support hybrid-36 either separately with our PDBReader.

(I have to say I don't feel very good about having our reader automatically trying to recognize every custom variation of the PDB format; perhaps we could have a flag like the old "strict=True" that actually fails when the format is not the true PDB one.... or at least a warning.)

[richardjgowers]

@orbeckst yeah maybe it would be better if you had to specify that this is required via keyword args. It might also make the resulting code cleaner/faster

[orbeckst]

I saw https://plumed.github.io/doc-v2.4/user-doc/html/pdbreader.html

Since PLUMED 2.4 we allow hybrid 36 format to be used to specify atom numbers. This format is not particularly widespread, but has the nice feature that it provides a one-to-one mapping between numbers up to approximately 80 millions and strings with 5 characters, plus it is backward compatible for numbers smaller than 100000. This is not true for notations like the hex notation exported by VMD.

Note the comment on VMD apparently doing its own thing vs following the hybrid-36 standard. (See also plumed/plumed2#226 )

What does packmol really do? VMD-style or hybrid-36?

I am all for supporting hybrid-36. Sample code: https://sourceforge.net/p/cctbx/code/HEAD/tree/trunk/iotbx/pdb/hybrid_36.py

[kain88-de]

Does it make sense to allow this format also for residue ids?