vdW bond distance is too large.

[electronicsbyjulie]

The most commonly cited figure is around 4A between atoms, depending on vdW radii. But going off of the chart on this page:

https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Physical_Properties_of_Matter/Atomic_and_Molecular_Properties/Intermolecular_Forces/Specific_Interactions/Van_Der_Waals_Interactions

We can see that the distance given for van der Waals binding is in agreement with the above, while that for hydrogen bonds is given as 3A (0.3nm). But this is the distance between heavy atoms, not the distance between hydrogen and hbond acceptor.

In olfaction, the majority of vdW forces are between C-H...H-C atoms. Consulting bindings.dat, the C-H bond is 1.09A long. Consulting elements.dat, the vdW radius of hydrogen is 1.1A. 1.09+1.1+1.09 = 3.28, in good agreement with the usual 4A-give-or-take figure. This plus the fact that vdW repulsion is due to vdW radii makes it seem like ~4A is the distance between heavy atoms, not between H atoms.

Further support for this interpretation can be found in Billesbolle (2022) in supplementary figure 6a showing the propionate ligand right up against residues of the binding pocket:

https://www.nature.com/articles/s41586-023-05798-y/figures/11

This figure does show carbon atoms at distances of more than 3A apart, but their hydrogen atoms are much closer together.

So I think the value in the distance column for vdW bonds in bindings.dat should be less than 4A.

[electronicsbyjulie]

Using the coordinates in RCSB 8f76.pdb, of the ligand position shown in the above image, atom C3 of the propionate is indeed a little less than 4A from the nearest atoms of Phe155 and Leu158, meaning the hydrogens are closer together.

ATOM    1186    CG  PHE A   155 123.911 116.152 85.579  1   35.73   C
ATOM    1187    CD1 PHE A   155 123.206 115.399 84.66   1   35.73   C
ATOM    1188    CD2 PHE A   155 124.186 115.594 86.814  1   35.73   C
ATOM    1189    CE1 PHE A   155 122.755 114.137 84.977  1   35.73   C
ATOM    1190    CE2 PHE A   155 123.746 114.322 87.123  1   35.73   C
ATOM    1191    CZ  PHE A   155 123.027 113.6   86.206  1   35.73   C
...                                         
ATOM    1215    CG  LEU A   158 123.469 117.023 81.018  1   42.24   C
ATOM    1216    CD1 LEU A   158 122.15  117.122 81.738  1   42.24   C
ATOM    1217    CD2 LEU A   158 123.571 115.737 80.23   1   42.24   C
...                                         
HETATM  8171    C1  PPI A   401 119.527 112.795 80.088  1   41.9    C
HETATM  8172    C2  PPI A   401 119.64  113.677 81.342  1   41.9    C
HETATM  8173    C3  PPI A   401 121.007 113.588 81.921  1   41.9    C
HETATM  8174    O1  PPI A   401 119.996 113.213 79.019  1   41.9    O
HETATM  8175    O2  PPI A   401 118.942 111.694 80.243  1   41.9    O

ATOM    1187    CD1 PHE A   155 2.20    1.81    2.74    3.95
ATOM    1189    CE1 PHE A   155 1.75    0.55    3.06    3.56

ATOM    1216    CD1 LEU A   158 1.14    3.53    -0.18   3.72
ATOM    1217    CD2 LEU A   158 2.56    2.15    -1.69   3.75

(Edited for formatting.)

[primaryodors]

Would be best to avoid reproducing copyrighted images. Not that there's anything desirable about journals putting scientific data behind paywalls, but why take the risk. This particular image is publicly available so I've edited your post to include a link instead.

That said, you raise a very good point.

[primaryodors]

There is now a vdW-bond-distance branch with the requested change. Please check out the branch and test it to make sure it passes unit tests, including a visual inspection of the output SDF and output dock files. If it looks good, go ahead and merge it into your working branch for #290.

[electronicsbyjulie]

Too close. The alcohols' carbons are 3.0A apart.

[electronicsbyjulie]

Also on PDB output the ASCII 3-letter bug seems to have returned.

[primaryodors]

When you fix the 3-letter bug, let's add a unit test to make sure it doesn't recur.