Superatoms/Generic groups in Marvin Sketch

User 3d9ae0af2d

29-03-2012 13:37:28

Hi all,


we use Marvin to prepare SDFs that will be used in an upcoming structure search feature at our website. Works well so far. The search logic is home-made using OpenBabel/PHP.


We do however have a few products in which part of the strucure is not disclosed (dye-labeled nucleic acids carrying dyes with a confidential structure). I would like to have SDFs of these products, too, to have them appear in the search results list.


But should I enter the confidential part? Marvin offers an "R" group which seems to represent a generic group. THis might work, but is too gerenal. I would rather use the abbreviation of the dye there, such as "ATTO565" or "DY485XL" etc. instead of a plain R.


So my question is if this can be done with Marvin and if the resulting SDFs conform to the standard and are suitable for structure searches by OpenBabel?


Thank you!


Best wishes,


Thomas.

ChemAxon d51151248d

30-03-2012 12:48:59

Dear Thomas !

The Marvin functionality that you need is an Alias atom for the confidental parts of your molecules. With these you can specify  a group for those parts of the molecule.


You can specify such an alias atom in MarvinSketch with the Periodic System > Advanced > Alias and add it to the molecule. You can use an alias for each of your dye molecules. After adding such aliases you can extract the molecule with Edit > Source in SDF format. SDF designates the line that shows an alias atom with 'A' and
the alias string in the next line.


The resulting structures should be handled with OpenBabel correctly. 

User 3d9ae0af2d

02-04-2012 11:11:58

Dear Daniel,


thank you for the answer! I tried it out with a quick test molecule, see a screen capture of the structure attached. Marvin automatically sub-scripts the number, but I guess this is just a cosmetic issue (can this be avoided somehow?).


The SDF looks okay, at least from my limited insight into SDF:




  Mrv0541 04021213052D          

  6  6  0  0  0  0            999 V2000
   -4.3902    5.4455    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
   -5.0576    4.9606    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
   -4.8027    4.1760    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
   -3.9777    4.1760    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
   -3.7228    4.9606    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
   -3.4928    3.5086    0.0000 C   0  0  0  0  0  0  0  0  0  0  0  0
  1  2  1  0  0  0  0
  1  5  1  0  0  0  0
  2  3  1  0  0  0  0
  3  4  1  0  0  0  0
  4  5  1  0  0  0  0
  4  6  1  0  0  0  0
A    6
DY485XL
M  END
$$$$



Why should I extract the molecule via Edit -> Source? I just save the structure as SDF from Marvin, and use this for our structure database. Is this okay?


Thanks again and best regards,


Thomas.

ChemAxon 5433b8e56b

02-04-2012 14:24:29

Hi Thomas,


Let me answer to you problem, you can avoud the automatic subscripting of numbers in the label when you type a "\n" before the numbers, and you will not need the Edit>Source if you get the molecule from the Sketcher via its API.


I hope this solution is suitable for you, currently there is not any general option, to turn off the automatic subscripting of numbers in labels, that i can suggest.


Regards,
Istvan

User 3d9ae0af2d

02-04-2012 14:34:09

thank you, Istvan, this has worked well, the numbers are normal now. I will try the resulting SDFs in our application


Is there a way to save pre-defined Aliases, or do we have to enter them each time again?


Best wishes,


Thomas.

ChemAxon 5433b8e56b

02-04-2012 14:41:00

If you want to reuse these then maybe the abreviated group is more suitable for you, which is a standard group also, and in this case you need to extend the abreviation group list, which is not as easy nowadays, but we are going to improve this in 5.10.


How you are createing these structures, are you using Marvin Applets, or MarvinBeans?


Istvan

User 3d9ae0af2d

02-04-2012 14:49:24

We're using MarvinBeans (the standalone Java application). If there is a way to save these abbreviation/aliases/whatever for later use with other structures, that would be great.


Best wishes,


Thomas.

ChemAxon 5433b8e56b

02-04-2012 15:07:59

In this case, if you do not want to wait until 5.10 then you can append the default abbreviation group file, but it needs some hacking in the MarvinBeans jars.


These are not signed jars, so you can freely edit it for example thru Total Commander or WinRar or such tools. You need to edit the current default abbrevgroup file from MarvinBeans-templates.jar inside the jar the path to the default abbravgroup file: cheamxon/marvin/templates/default.abbrevgroup.


You can find documentation on the file format on this page.


I hope this helps, in 5.10 it would be easier, currently we are working on this area.


Istvan

User 3d9ae0af2d

03-04-2012 08:51:58

Hi Istvan,


I guess I will wait for 5.10. Do you have any timeline when this will version be published?


Best wishes,


Thomas.

ChemAxon f052bdfe3c

04-04-2012 09:44:49

Hi Thomas,


The 5.10 is planned to be released in early June.


Best Regards,


Efi

User 3d9ae0af2d

04-04-2012 09:47:20

thank you, Efi! This sounds good.


Best regards,


Thomas.

User 3d9ae0af2d

13-04-2012 14:57:32

Hi all,


well, it seems that OpenBabel has a problem with the Alias-containing SDFs. We drew a few that way, and they are processed by OpenBabel now to generate fingerprints for structure search. The SDF looks okay from my point of view, and looks okay in Marvin (see below). However, the image that OpenBabel generates is corrupt. See attached the SDF, and below the image. I cannot evaluate if the other data such as SMILES are okay, though.


Best regards,


Thomas.

ChemAxon f052bdfe3c

18-04-2012 19:29:23

May I ask that the fingerprints generated by OpenBabel  are also corrupt?

User 3d9ae0af2d

14-05-2012 13:06:07

Hello,


sorry for the long delay, I was on vacation and got ill afterwards...


Are you sure THAT the fingerprints are not okay, or are you just guessing that this yould be a potential cause?


Meanwhile I found out that we use mol2ps (http://merian.pch.univie.ac.at/~nhaider/cheminf/mol2ps.html) for generation of the images.This could also be a reason - the SDFs are fine, but mol2ps doesn't interpret them correctly.


Best regards,


Thomas.

ChemAxon f052bdfe3c

21-05-2012 16:16:17

Hi,


I was just guessing, because you wrote  "However, the image that OpenBabel generates is corrupt.


So, does it seem now  that mol2ps caused the problem?


Best Regards,


Efi

User 3d9ae0af2d

22-05-2012 08:58:56

Hi Efi,


exactly, it might be mol2ps, but I haven't gone into depth with this so far. We also have to test if OpenBabel is working correctly with these files. This might take some time.


Best regards,


Thomas.

ChemAxon f052bdfe3c

24-05-2012 13:27:59

There Thomas,


 


Please let me know if you find the root of the problem. I am really interested in it.


Thank you,


Efi

User 3d9ae0af2d

24-05-2012 13:30:18

Sure, Efi - will do!


Best regards,


Thomas.

User 3d9ae0af2d

21-06-2012 09:19:19

Hi all,


I'm in the process to dig a little bit deeper into this problem. I seems that the SDFs created by Marvin are correct (and work okay in our OpenBabel based search engine), but mol2ps/mol2png (which we use to generate images from the structure files) can't cope with some of these features. Fortunately, Norbert Haider from Vienna, the developer of mol2ps, is extremely helpful to fix this.


One question regarding the SDFs with generics: attached is an SDF of an immobilized ATP derivative. The stationary phase is represented by "R", which displays correctly in Marvin (see screenshot). However, when you look in the SDF, there is "R#" instead of "R" - and mol2ps will consequently label this group "R#".


-19.4669    0.7610    0.0000 R#  0  0  0  0  0  0  0  0  0  0  0  0


Is there a reason why Marvin adds the # sign to the "R" group?


Thank you + best regards,


Thomas.

User 3d9ae0af2d

21-06-2012 09:47:37










ifajth wrote:

In this case, if you do not want to wait until 5.10 then you can append the default abbreviation group file, but it needs some hacking in the MarvinBeans jars.



Hello Istvan,


since 5.10 is out now, is there a way to save custom labels? I have checked Periodic System -> Advanced, but this page looks just as did before.


Thank you!


Best regards,


Thomas.

ChemAxon 5433b8e56b

22-06-2012 13:21:07

Hi,


there is a documentation on how to extend the abbreviated group list, when the groups are provided they will be available in the sketcher by typing. Link for the docs: http://www.chemaxon.com/marvin/help/sketch/sketch-basic.html#User_defined_abbreviated_groups.


For the R# problem in export I am forwarding your question to the Core team they are dealing with structure import/export issues. They will answer soon.


Regards,
Istvan

User 3d9ae0af2d

22-06-2012 13:28:22

Hello Istvan,


many thanks! We will have a look at the documentation.


Best wishes,


Thomas.

ChemAxon 25dcd765a3

22-06-2012 14:28:54

Dear Thomas,



One question regarding the SDFs with generics:
attached is an SDF of an immobilized ATP derivative. The stationary
phase is represented by "R", which displays correctly in Marvin (see
screenshot). However, when you look in the SDF, there is "R#" instead of
"R" - and mol2ps will consequently label this group "R#".


-19.4669    0.7610    0.0000 R#  0  0  0  0  0  0  0  0  0  0  0  0


Is there a reason why Marvin adds the # sign to the "R" group?



I have tried to convert the attached AC-101.sdf to sdf again actually it is exported as


-19.4669    0.7610    0.0000 R   0  0  0  0  0  0  0  0  0  0  0  0

so the # disappeared.


This is what I run:


molconvert sdf AC-101.sdf

Which Marvin version are you using? How did you get the AC-101.sdf?


All the best


Andras

User 3d9ae0af2d

22-06-2012 14:48:30

Hi Andras,


the file was created with a previous version of Marvin Sketch a while ago. Currently we use the latest one (5.10).


I will try to save the files in question (which are a few) with the latest version and will see if this helps.


Thank you!


Best wishes,


Thomas.

User 3d9ae0af2d

25-06-2012 08:40:30

Hi Andras,


just a quick thought - before running molconvert on all our SDFs, can I just replace R# by R in the SDFs in question using a text editor?


Best wishes,


Thomas.

ChemAxon d26931946c

25-06-2012 13:41:34

Hi Thomas,


 


We've fixed this export issue (writing R# instead of R) in 5.9.0.


If you remove the # the import works fine. Since 5.9 we ignore that character.


 


Regards,


Peter


 

User 3d9ae0af2d

25-06-2012 13:48:34

Hi Peter,


thanks - we use 5.10 from now on. I just wonder if we can fix the SDFs containing R# quickly by using a text editor (changing R# to R), or if we have to run molconvert on all these files (or open and save them with the recent Marvin version)...


Best wishes,


Thomas.

ChemAxon d26931946c

25-06-2012 14:21:49

Hi, 


You can replace R# with R. It will not effect our import in 5.10.


Just be sure that don't change the number of whitespaces.


Regards,


Peter

User 3d9ae0af2d

25-06-2012 14:49:44

Hi Peter,


thanks a lot, that will help!


Best regards,


Thomas.