mk_export.py Runtime error

[eightmm]

Hi! I tried to convert dlg, the result file of AutoDock-GPU, to sdf by using mk_export.py. At this time, the following error was output.

Looking at DLG and PDBQT, I found that there are 4 hydrogen atoms.

Is this problem caused by rdkit not recognizing hydrogen properly? or is there another reason?

Thanks!

[diogomart]

Hi! The explicit H in the REMARK SMILES line are unusual, my first guess is that the error is relates to that. Could you post the input ligand pdbqt as text so i can take a look tomorrow?

[eightmm]

@diogomart

REMARK SMILES [H]c1c(N([H])[H])c([H])c2nc3c([H])c(N([H])[H])c([H])c([H])c3c([H])c2c1[H] REMARK SMILES IDX 26 1 2 2 3 3 7 4 12 5 14 6 18 7 20 8 23 9 10 10 9 11 22 12 REMARK SMILES IDX 25 13 11 14 4 15 5 16 6 17 15 18 16 19 17 20 REMARK H PARENT REMARK Flexibility Score: inf ROOT ATOM 1 C UNL 1 8.982 23.182 49.516 1.00 0.00 0.012 A ATOM 2 C UNL 1 9.670 23.649 48.475 1.00 0.00 0.026 A ATOM 3 C UNL 1 10.665 22.851 47.794 1.00 0.00 0.034 A ATOM 4 C UNL 1 10.902 21.573 48.264 1.00 0.00 0.054 A ATOM 5 C UNL 1 9.939 17.864 51.143 1.00 0.00 0.054 A ATOM 6 C UNL 1 9.231 17.253 52.163 1.00 0.00 0.034 A ATOM 7 C UNL 1 8.222 18.030 52.852 1.00 0.00 0.026 A ATOM 8 C UNL 1 7.968 19.299 52.528 1.00 0.00 0.012 A ATOM 9 C UNL 1 8.475 21.272 51.058 1.00 0.00 0.020 A ATOM 10 N UNL 1 10.385 19.776 49.768 1.00 0.00 -0.248 NA ATOM 11 C UNL 1 10.196 21.070 49.336 1.00 0.00 0.073 A ATOM 12 C UNL 1 8.690 19.965 51.464 1.00 0.00 0.001 A ATOM 13 C UNL 1 9.190 21.846 50.020 1.00 0.00 0.001 A ATOM 14 C UNL 1 9.691 19.173 50.793 1.00 0.00 0.073 A ENDROOT BRANCH 3 15 ATOM 15 N UNL 1 11.274 23.326 46.684 1.00 0.00 -0.399 N ATOM 16 H UNL 1 11.077 24.283 46.358 1.00 0.00 0.156 HD ATOM 17 H UNL 1 11.935 22.731 46.164 1.00 0.00 0.156 HD ENDBRANCH 3 15 BRANCH 6 18 ATOM 18 N UNL 1 9.453 15.962 52.502 1.00 0.00 -0.399 N ATOM 19 H UNL 1 10.173 15.416 52.007 1.00 0.00 0.156 HD ATOM 20 H UNL 1 8.902 15.525 53.255 1.00 0.00 0.156 HD ENDBRANCH 6 18 TORSDOF 2

This pdbqt file is the input pdbqt.

The above pdbqt file uses "mk_prepare_ligand.py" to change this sdf file.

Thanks!

[diogomart]

Hi again, Could you post the SDF file as a block of code? It needs to be formatted as code so the spaces are preserved. To format text as code use triple backticks before and after the code block. https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax#quoting-code

[eightmm]

1bcu_ligand

Created by X-TOOL on Fri Sep 26 17:34:14 2014
 27 29  0  0  0  0  0  0  0  0999 V2000
    8.9820   23.1820   49.5160  C 0  0  0  2  0  3
    9.6700   23.6490   48.4750  C 0  0  0  2  0  3
   10.6650   22.8510   47.7940  C 0  0  0  1  0  3
   10.9020   21.5730   48.2640  C 0  0  0  2  0  3
    9.9390   17.8640   51.1430  C 0  0  0  2  0  3
    9.2310   17.2530   52.1630  C 0  0  0  1  0  3
    8.2220   18.0300   52.8520  C 0  0  0  2  0  3
    7.9680   19.2990   52.5280  C 0  0  0  2  0  3
    8.4750   21.2720   51.0580  C 0  0  0  2  0  3
   10.3850   19.7760   49.7680  N 0  0  0  1  0  2
   10.1960   21.0700   49.3360  C 0  0  0  1  0  3
    8.6900   19.9650   51.4640  C 0  0  0  1  0  3
    9.1900   21.8460   50.0200  C 0  0  0  1  0  3
    9.6910   19.1730   50.7930  C 0  0  0  1  0  3
   11.2740   23.3260   46.6840  N 0  0  0  3  0  3
    9.4530   15.9620   52.5020  N 0  0  0  3  0  3
    8.2489   23.8205   49.9955  H 0  0  0  1  0  1
    9.4780   24.6579   48.1280  H 0  0  0  1  0  1
   11.6536   20.9575   47.7829  H 0  0  0  1  0  1
   10.7001   17.3042   50.6115  H 0  0  0  1  0  1
    7.6572   17.5677   53.6535  H 0  0  0  1  0  1
    7.2050   19.8431   53.0728  H 0  0  0  1  0  1
    7.7239   21.8637   51.5686  H 0  0  0  1  0  1
   11.0771   24.2830   46.3581  H 0  0  0  1  0  1
   11.9348   22.7308   46.1644  H 0  0  0  1  0  1
   10.1726   15.4160   52.0071  H 0  0  0  1  0  1
    8.9024   15.5253   53.2550  H 0  0  0  1  0  1
  1 13  4  0  0  1
  1  2  4  0  0  1
  2  3  4  0  0  1
  3 15  1  0  0  2
  3  4  4  0  0  1
  4 11  4  0  0  1
 11 10  4  0  0  1
 11 13  4  0  0  1
 13  9  4  0  0  1
  9 12  4  0  0  1
 12  8  4  0  0  1
 12 14  4  0  0  1
 14  5  4  0  0  1
 14 10  4  0  0  1
  5  6  4  0  0  1
  6 16  1  0  0  2
  6  7  4  0  0  1
  7  8  4  0  0  1
  1 17  1  0  0  2
  2 18  1  0  0  2
  4 19  1  0  0  2
  5 20  1  0  0  2
  7 21  1  0  0  2
  8 22  1  0  0  2
  9 23  1  0  0  2
 15 24  1  0  0  2
 15 25  1  0  0  2
 16 26  1  0  0  2
 16 27  1  0  0  2
M  END
> <MOLECULAR_FORMULA>
C13H11N3

> <MOLECULAR_WEIGHT>
209.2

> <NUM_HB_ATOMS>
3

> <NUM_ROTOR>
0

> <XLOGP2>
1.99

$$$$

Here!

[diogomart]

It looks like the hydrogen count field of the MOL block is used (page 13 in the specification), and RDKit does not remove hydrogens. I don't know if this is the expected RDKit behavior, but meeko should be checking that those hydrogens were removed, so I think it needs fixing on our end.

[diogomart]

So, it turns out that the input SDF has the HCount field set for the hydrogens, for example for the last atom:

    8.9024   15.5253   53.2550  H 0  0  0  1  0  1
                                           ^
                                           |
                                           hcount

According to the specification hcount = 1 sets the number of implicit Hs to zero. But because the HCount field is specified (i.e., has a non-zero value), RDKit adds a query and does not remove the atom by default.

Based on the header it looks like the SD file was written by X-TOOL. Do you know why the hcount fields were set?

[eightmm]

Sorry, I don't know . I just tested it with the files in the PDBbind dataset.

[diogomart]

Thanks! The easy fix on meeko's end is to tell RDKit to remove Hs even if they have hcount queries, but I was concerned that those queries were added on purpose. Glad to hear that they weren't :-)

[diogomart]

fixed in ba15b7dc41684405eeebd4b66d6e4a655a9178f2 Thanks for reporting this!