openmolecules.org

 
Home » DataWarrior » Bug Reports » Faulty SDF import (Incorrect determination of bonds)
Faulty SDF import [message #1087] Mon, 19 October 2020 11:30 Go to next message
richards99 is currently offline  richards99
Messages: 42
Registered: May 2020
Location: UK
Member
Hi,
There are still issues with regards the import of SDF files and its interpretation of the bonding.
This time it is not down to specified aromatic bonds but simply kekulised bonds.

Take for instance this SMILES string of a protonated heterocycle.
CNC1=C2C=C3OCCOC3=CC2=[NH+]C=C1

When this is saved as an SDF file using Marvin for instance, and then imported into DataWarrior the bonding is completely wrong.
I have attached the SDF of this smiles string above.

The output it provides is attached
index.php?t=getfile&id=277&private=0

Re: Faulty SDF import [message #1088 is a reply to message #1087] Mon, 19 October 2020 11:32 Go to previous messageGo to next message
richards99 is currently offline  richards99
Messages: 42
Registered: May 2020
Location: UK
Member
I am also using the dev version of DataWarrior where you mentioned there were fixes regard SDF importing, but the issue remains.
Re: Faulty SDF import [message #1089 is a reply to message #1088] Wed, 21 October 2020 09:15 Go to previous messageGo to next message
nbehrnd is currently offline  nbehrnd
Messages: 204
Registered: June 2019
Senior Member
My observation differs:

1) First I pasted the SMILES string into DW (5.2.1, natively Linux version [used in Debian 10 / unstable]) and obtained a pyridinium:

index.php?t=getfile&id=282&private=0

2) Then the .sdf provided by you (as generated by Marvin) was read by a separate, newly started instance of DW, again yielding a pyridinium:

index.php?t=getfile&id=284&private=0

which seems chemically plausible, matches (by constitution) both the former and the visual output when requesting OpenBabel (3.1.0) to visualize the .sdf as a .png:

index.php?t=getfile&id=285&private=0
Re: Faulty SDF import [message #1090 is a reply to message #1089] Wed, 21 October 2020 11:54 Go to previous messageGo to next message
richards99 is currently offline  richards99
Messages: 42
Registered: May 2020
Location: UK
Member
It seems the standard download of DW does display it correctly, but the DEV release of DW is where the structure is displayed incorrectly.
This post ( http://openmolecules.org/forum/index.php?t=msg&th=385&am p;start=0&)mentions SDF importing has been fixed in the DEV release, but I think this has introduced a new error.

[Updated on: Wed, 21 October 2020 11:57]

Report message to a moderator

Re: Faulty SDF import [message #1092 is a reply to message #1090] Thu, 22 October 2020 12:28 Go to previous messageGo to next message
nbehrnd is currently offline  nbehrnd
Messages: 204
Registered: June 2019
Senior Member
Affirmative. The pristine installation of DW for Linux using the stable release (last edit 23 Feb 2020) loads your .sdf file and then displays the single protonated pyridinium, as described in my initial post. But the substitution of elder / original datawarrior.jar file by the one obtained from the development branch for Linux (1), deposit into the /opt/datawarrior directory (last edit 13 Sep 2020), then yields the questionable double-protonation on the pyridine; which matches 1:1 your initial observation.

Because rebuilding DW today (22 Okt 2020) from its public sources (2, last edit 13 Sep 2020) in Linux equally yields an executable displaying again the erroneous «double pyridine protonation», possibly the issue spreads out to the development version for Windows,(3) too.

Norwid

(1) http://www.openmolecules.org/datawarrior/dw521x.zip
(2) http://www.openmolecules.org/datawarrior/dw521source.tar.gz
(3) http://www.openmolecules.org/datawarrior/dw521win.zip
Re: Faulty SDF import [message #1095 is a reply to message #1092] Fri, 23 October 2020 10:13 Go to previous message
thomas is currently offline  thomas
Messages: 646
Registered: June 2014
Senior Member
Dear All,

there has been a bug in OpenChemLib regarding open valences of charge atoms in certain situations, which seems to have caused the issue. DataWarrior stores structures in a canonical representation. Therefore, it needs to resolve delocalized bonds into alternating single/double bonds when displaying. Because of the valence bug the charged nitrogen was not seen as having the needed open valence. The bug was fixed Sep. 30 in OpenChemLib, but took till yesterday to materialize in a DataWarrior build.

Thank you all for reporting.

Thomas

[Updated on: Fri, 23 October 2020 10:14]

Report message to a moderator

Previous Topic: Export of Chiral compounds
Next Topic: 3D structure display is lost after hovering over certain structures
Goto Forum:
  


Current Time: Thu Mar 28 19:19:13 CET 2024

Total time taken to generate the page: 0.08473 seconds