mol2 format problems

Hi!




I'm using the latest Marvin version, and I need a correct mol2 file as output. The probel is that I cannot write the charges and the atom names for the carbonyl group are not correct. It cannot be readed by Sybyl.




Is there a simple way to have all datas correctly in Marvin? Or which other format saves the charges?






Annamaria

Hi Annamaria!








I'm afraid the present mol2 implementation does not yet support charges. It is scheduled for implementation in version 5.3 - that will be released in Q2 2009.


The most comprehensive file format Marvin supports is the Marvin Document format, MRV. That can store partial charges, however, it is not compatible with other vendor tools.


Another feasible format MDL Mol that can represent partial atomic charges, though I'm not sure that the application you are trying to use will accept MDL Mol instead of Tripos mol2. It's worth giving it a try.





Regards


Miklos

Hi Miklos!





Thanks for your response. Q2 2009 means which month exactly? The mrv format is maybe not the best solution in my case, because I'm not using only Marvin.


How can I exactly get the charges in the output file? Only by using the charge module in cxcalc?





Annamaria

Hi!





I tried to convert my molecules via the molconvert option, but I have the following error message: (for molconvert sdf *.mol2)


82: error: Molecule cannot be imported


The molecule is saved in mol2 format using Marvin...





What do I wrong?

Hi Panni,





1. Q2 means June :)





2. Indeed, first you need to calculate the charges in order to safe them in your output file.





3. Regarding the molconverter problem, that looks serious. The command line looks perfect. Can you send as your mol2 file (created by marvin)? - if structures are not confidential, of course. That would help us trace back the problem. Perhaps your original structure (in SDF? or what format) would be even more useful.





Thanks


Miklos

Hi Miklos,

Any chance the mol2 partial charges export is implemented already? I could really use it right now...

Cheers,

Nicolas

Hi Nicolas,

We scheduled a task for the release 5.2.3 to improve mol2 import/export. We can generate mol2 atom types in export, though partial charges can only be calculated with our charge plugin. That means that figures will not agree with the native Sybyl charges (though I hope signs at least will :-) )

Would such partial solution acceptable for you?

Regards,

Miklos

 

Hi Miklos,

I'm afraid the atom types are not enough, partial charges are necessary for docking programs.

From
what I understand, the charge plugin can only display or tag atoms in
the mrv format. I'm not sure of what you mean by "that means that
figures will not agree with the native Sybyl charges (though I hope
signs at least will :-) )"

I've
circumvented the problem by using openbabel, which can also split sdf
files into individual mol2 files (any plans for that?). It would be
great if chemaxon tools could do all the work (conformation
calculations are still much faster!)

Cheers

Nicolas

 

 

 

Hi Nicolas,

- I certainly understand that charges are important and we will provide partial atomic charges. We will use our existing charge plugin to calculate charges and those will be stored in the mol2 file as required.

- We cannot reproduce the exact same calculation what Sybyl carries out when estimating the charges. Their model and our charge model are certainly different, so there will be differences in the charge values. However, I do not expect major discrepancies. 

Also consider, that MOL2 supports various charge models: NO_CHARGES, DEL_RE, GASTEIGER, GAST_HUCK, HUCKEL, PULLMAN, GAUSS80_CHARGES, AMPAC_CHARGES, MULLIKEN_CHARGES, DICT_ CHARGES, MMFF94_CHARGES, USER_CHARGES.

We will use USER_CHARGES, though our model is an extension of Mulliken.

- I admit that splitting into separate files is a very handy feature, I will create an internal task for that feature request, though I cannot tell when this will be delivered, it it will be implemented at all, it's beyond my authority.

- When you say that "conformation calculations are still much faster" you mean that in molconvert these are faster than in Babel?

Kind regards,

Miklos

 

Hi Miklos,

I was confused about what you said. I didn't expect you to reproduce the Sybyl charges... any coherent charges would be okay (I got your sign joke), and it seems your calculations based on electronegativity does a good job.

About openbabel: it calculates GASTEIGER charges, but the 3D generation takes about 2000 compounds/day (and got stuck after 4500 molecules in my latest run, I am re-running now) whereas molconvert did about 9000 molecules in a few hours.

Yes, splitting is nice ;-)

Cheers,

Nicolas

Hi Nicolas,

Thank you for your response. Glad to learn that our charges are feasible, and in particular that the 3D generation in marvin/molconverter works faster than Babel.

Splitting has been implemented since then and it will be released in version 5.2.2. along with the mol2 export.

Regards,

Miklos

 

Hi!

I'm testing the new mol2 format from last release, it is really nice, thanks :).

There are only a little technical problem in the charge values: in the original mol2 format, the charges should be given given with 4 floats, like

0.1234

This part is not ok in 2 cases: eg.

4       C4     0.0000     0.0000     0.0000     C.2     1       noname  -9.0E-4         

 I have in this case -9E-4 instead of -0.0009

and

7       C7     0.0000     0.0000     0.0000     C.2     1       noname  0.063

0.063 instead of 0.0630

It is only a little little output problem.

Thanks,

Panni

 

 

Dear Panni,

 

Thanks for the testing and the bug report. The fix will be included in the forthcomming 5.2.4 release expected in a few days.

 

Regards,

Gabor