Francois,

There is not a direct unambiguous physical observable that one can point to
that defines a van der Waals radius; they are looked up in tables.
The RDKit and OpenBabel are clearly using different tables.

As to which to use: there isn't really an answer to the question; it comes
down to what you're trying to do and which table you choose to use. If you
just want a vdW radius (and don't care about a vdW distance for a
particular pair of atoms) you can just use a single tabulated value like
what you give below. Given that I no longer remember where the RDKit value
in atomic_data.cpp comes from and the OpenBabel one likely has a reference
attached to it: I'd take the OpenBabel one.

If you care more about what a particular vdW contact length should be,
you're probably better off using a more sophisticated approach. Paolo has
some information on that in this thread:
https://sourceforge.net/p/rdkit/mailman/message/35428577/

-greg



On Mon, Jun 19, 2017 at 7:49 AM, Francois BERENGER <
beren...@bioreg.kyushu-u.ac.jp> wrote:

> Hello,
>
> Sometimes, as a computer scientist, I am quite worried by chemical
> software libraries:
>
> $ cat data/ethanol.pqr
> COMPND    ethanol
> AUTHOR    GENERATED BY OPEN BABEL 2.3.2
> HETATM    1  C   LIG     1      -0.017  -0.601   0.000  0.04138432 1.700  C
> HETATM    2  C   LIG     1       1.247   0.248  -0.000 -0.04182526 1.700  C
> HETATM    3  H   LIG     1       1.285   0.894  -0.888  0.02516417 1.100  H
> HETATM    4  H   LIG     1       1.270   0.916   0.872  0.02516417 1.100  H
> HETATM    5  H   LIG     1       2.150  -0.369   0.015  0.02516417 1.100  H
> HETATM    6  H   LIG     1      -0.096  -1.231  -0.903  0.05542120 1.100  H
> HETATM    7  H   LIG     1      -0.096  -1.231   0.903  0.05542120 1.100  H
> HETATM    8  O   LIG     1      -1.086   0.337   0.000 -0.39527743 1.520  O
> HETATM    9  H   LIG     1      -1.944  -0.125  -0.003  0.20938346 1.100  H
> CONECT    1    8    7    2    6
> CONECT    1
> CONECT    2    1    5    4    3
> CONECT    2
> CONECT    3    2
> CONECT    4    2
> CONECT    5    2
> CONECT    6    1
> CONECT    7    1
> CONECT    8    1    9
> CONECT    9    8
> MASTER        0    0    0    0    0    0    0    0    9    0    9    0
> END
>
> # cat data/ethanol.plr
> COMPND ethanol
> -0.016600 -0.601400 0.000000 -0.203500 1.950000
> 1.247400 0.248500 -0.000100 0.144100 1.950000
> 1.285200 0.894300 -0.887700 0.123000 1.200000
> 1.270100 0.915700 0.872200 0.123000 1.200000
> 2.149700 -0.369300 0.015500 0.123000 1.200000
> -0.095800 -1.230700 -0.903200 0.123000 1.200000
> -0.095700 -1.230800 0.903400 0.123000 1.200000
> -1.085800 0.337100 0.000300 -0.289300 1.700000
> -1.944300 -0.124700 -0.003200 -0.267700 1.200000
> END
>
> The .plr file was generated with rdkit.
> The .pqr file with obabel.
>
> Maybe I'm out of luck, or I don't understand something, but the two files
> don't agree
> on any single VdW radius (5th column in the .plr file; 10th column in the
> .pqr
> file for HETATM lines).
>
> Here is the code I used for rdkit:
> ---
> periodic_table = Chem.GetPeriodicTable()
>
> def get_radius(atom):
>     return periodic_table.GetRvdw(atom.GetAtomicNum())
>
> [...]
> atoms = mol.GetAtoms()
> [...]
> radius = get_radius(atom)
> [...]
> ---
>
> Am I doing something wrong?
>
> Isn't there a more direct way to access the VdW radius of an atom
> when iterating over the atoms of a molecule?
>
> Who should I trust? obabel? rdkit? the nearest real-world chemist?
>
> Thanks,
> F.
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Rdkit-discuss mailing list
> Rdkit-discuss@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Rdkit-discuss mailing list
Rdkit-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss

Reply via email to