Can Fragmenter get same UID while different fragment?

HI,


I have a qustion about the fragmenter.That is when I use fragmenter to


split molecules ,I got the same UID but obviously these two fragments are different, How to explain that? Can anyone help me?


Thank you!

I move your post to the Fragmenter forum. Next time, please posts only Marvin releated questions to the Marvin forum.

Thank you.

Tamas wrote:

I move your post to the Fragmenter forum. Next time, please posts only Marvin releated questions to the Marvin forum.

Hi,





I checked fragmenter and it work all right. Though I am not aware of your cleavage rules and other parameter settings that may cause different behaviour.


Perhaps, as a first step, you may wish to try some of the examples provided in the JChem User Manual and in the JChem examples. The simples way to do some basic test is to go to the examples folder in your JChem installation directory. There you find a fragmenter folder with numerous examples along with a short guide: index.html.


The first sample command in that guide is

Code:

fragment -c Fragmenter.xml input.mol -f sdf:-a -o fragments1.sdf

After it successfully completed running, use mview to display the output file fragments1.sdf. Check UIDs and compare them against the documentation. Is there any difference or any suspicious UIDs?





If this example works all right and in your particular case UID are still wrong then we need more specific information about your actual set-up (ie. configuration file, input molecules etc). I hope those are not confidential though.





Does this help?





Regards,


Miklos

Thank you for your reply. But It's my fault that I didn't take an example.


Two sdf files are listed in attanchments.


Then I launch a command window,


and press two comands





fragment -c FragmenterRecap.xml 1.sdf -f sdf:-a -o 1_splited.sdf


fragment -c FragmenterRecap.xml 2.sdf -f sdf:-a -o 2_splited.sdf





Then I get two fragments files.


But You can see that , the UID of No.3 fragment in 1_splited.sdf is same as NO.1 fragments in 2_splited.sdf, Obviously, they are not same in chemical formula.


That is all.

mvargyas wrote:

Hi, I checked fragmenter and it work all right. Though I am not aware of your cleavage rules and other parameter settings that may cause different behaviour. Perhaps, as a first step, you may wish to try some of the examples provided in the JChem User Manual and in the JChem examples. The simples way to do some basic test is to go to the examples folder in your JChem installation directory. There you find a fragmenter folder with numerous examples along with a short guide: index.html. The first sample command in that guide is Code: fragment -c Fragmenter.xml input.mol -f sdf:-a -o fragments1.sdf After it successfully completed running, use mview to display the output file fragments1.sdf. Check UIDs and compare them against the documentation. Is there any difference or any suspicious UIDs? If this example works all right and in your particular case UID are still wrong then we need more specific information about your actual set-up (ie. configuration file, input molecules etc). I hope those are not confidential though. Does this help? Regards, Miklos

Hi,





thank you for your response and the detailed description of the situation along with the related files. It certainly helped me understand what went wrong.





The UID that Fragmenter assigns to the fragments is not a unique identifier in an absolute sense, that is, it is not something like a 'canonical code' of that fragment. Instead, UID is relative and specific to the actual run of Fragmenter and has no connection to other execution of it. This is why you experienced that different fragments in different runs get the same id.





I hope I managed to clarify the situation. Please let me know if any further explanation is needed.





If you need to distinguish between different fragments generated in independent program executions than you need to assign a canonical code to your fragments (e.g. canonical smiles, Morgan name etc.).


Alternatively, you may consider to merge your separate input files into one SDFile and apply Fragmenter to that single file.





I hope it helps.





Regards,


Miklos

Thank you for your help .


My idea is that if the UID can't special every fragment,why should we call it as UID? :) Why not cannonize them?


In fact I have tried to merge the two files into one, then I got the same result(same UID).


Anyway,thank you very much, Your group do a lot of perfect work.Have a nice day!





:)





Lirong

Hi Liring,





apologies for getting back so late, we are very busy with organizing our UGM.


What you experienced with the merged files really surprises me so i'll check that later when ugm is over. I'll let you know.





Miklos

I assume you mean the "carbon-carbon:1" sign in the atom label.
This ID identifies the cleavage rule with atom map and not the fragment molecule, see the FragmenterRecap.xml config file:

<Action ID="carbon-carbon" Structure="[C:1]-[C:1]>>[C:1].[C:1]"/>

 

FragmentSatistics uses a coded version of these IDs for each atom as well as the unique SMILES string to identify fragments.

Some new fragmentation options will be available in the 5.3 release:

• Cleave single bonds connecting chain and ring atoms (with few special cases)
  


• Do not cleave bonds of terminal atoms
  


• Generate elemental fragments only, no permutations
  


• Label connection points with any atoms or Al and Ar atoms
  

 

Documentation:

• Fragmenter
  


• FragmentSatistics
  


• working examples