Problems saving as *.mol file

Hi,

we need to have some planar molecules as *.mol files as input for quantumchemical pi-SCF calculations. And the planar molecules should be placed in the xy-plane.

But there is a problem in saving 2D-structures in *.mol files, see the attached screenshot.

Doing the following

1) Putting naphthalin on the canvas

2) Add - Add Explicit Hydrogens

3) Clean 2D

produces the planar molecule on the canvas.

Saving this structure as *.xyz file gives reasonable coordinates (upper screenshot part).

But saving as *.mol produces the cordinates multiplied by the factor 0.536 ???

So we cannot use those files as input files.

Using Clean 3D produces a molecule which is no more in the xy-plane. It's planar of course, but we need to have it in the xy-plane.

And doing Clean 3D and then Clean 2D produces the molecule back in the xy-plane (with better C-H bonds 1.09 A) but saving it in 2D as *.mol file still gives the wrong coordinates multiplied by the factor 0.536.

What's wrong ?

Best regards, Hans-Ulrich

Hi Hans-Ulrich,

I moved this topic to the Structure representation forum.

Best regards,

Akos

Hi Akos,

I don't agree. Saving a molecule with inacceptable coordinates is a bug of MarvinSketch and not of structure representation.

Best regards, Hans-Ulrich

Hi Hans-Ulrich,

This forum also deals with import/export from/to molecule file formats.

Best regards,

Akos

Hi Hans-Ulrich,

It isn't a bug, it is the normal function. In MDL Molfile v2000 there is a multiplication. If you want to got a molfile with the origial coordinates, use extended molfile (v3000).

Or, the other solution is using molconvert to your file:

molconvert mol:b1.54 original.mol -o result.mol

I hope it helps.

Janos

Hi Janos,

OK, it's not a bug. But what is the purpose to multiply the 2D-coordinates ? Saving a 3D structure in a not "Extended" mol file does not multiply the coordinates.

And what is the value of the multiply factor ?

And the "Extended Mol File Format" is unusable for us. The format is totally different, see attached image. So the quantum chemical program cannot read it. And the extender *.mol is the same ! How to now which format is in the file ?

And to use the command line program molconvert is inappropriate in student courses and in a teaching computer pool.

We have a "workaround". But before implementing and showing it here, I must have the answers to my questions here.

Best regards, Hans-Ulrich

Hi Hans-Ulrich,

The default scale constant is 0.825, this way we convert the bondlengths to ISISDraw units. We set the default bond length to 1.54 in Marvin 3.5 (2004). At this moment we started to rescale the coordinates in MDLMol export, probably for ISIS compatibility.

The rescale only appears in 2D molecules.

In the 4th line of the *mol file there is a version identifier that tells the format of the file (v2000 or v3000).

 

Best regards,

Peter

Hi Peter,

according to your information I changed the program reading the input.

When it finds in the 2nd line the letters "2D" then the input coordinates are multiplied by 1.86667. This factor is coming from the rescale 1.54 / 0.825.

Now the program works as it should do it and recognizes the correct topologies. Even the very long C-H distances in the 2D mol-files are now identified.

Thanks for your hints.

Best regards, Hans-Ulrich

PS: In FORTRAN 77 it's not so easy to analyse strings ...