How regeneration works?

Hi!

I could not find the answer in Chemaxon documentation how regeneration works, so I am posting the question here. The main problem - we have structures in our database with strange SMILES like:
Nc1nc(cs1)C12CC3CC(CC(C3)C1)C2 |THB:11:10:7:13.12.14,11:12:9.10.15:7,14:12:9:15.6.7,14:6:9:13.11.12|,
S=C=NC12CC3CC(CC(C3)C1)C2 |TLB:2:3:6:10.8.9,THB:8:7:4:10.9.11,8:9:6.7.12:4,11:9:6:12.3.4,11:3:6:10.8.9|,
etc.
As these molecules have low CD_ID values, we believe they were imported with old Chemaxon versions (v4).

Questions:
1) what does the THB, TLB stands for? is it normal from v5 perspective?
2) is regeneration modifying SMILES?
3) if the ending part is incorrect, can it be eliminated with regeneration? if yes, how regenerator must be run (as default regeneration when Chemaxon is upgraded does not modify SMILES);
4) in Chemaxon documentation http://www.chemaxon.com/jchem/doc/dev/java/api/index.html?chemaxon/jchem/db/RegenerationChecker.html I found constants which I believeare used in regeneration process;
Is there any documentation which describes the purpose of these constants and how these constants are used? Can we pass regeneretion type to regnerator via switch, configuration file, etc.?

Thanks for answer in advance.

Hi,

We use an extension of the SMILES format called cxsmiles to be able to represent all structurally important features for internal use by our search process.

Detailed description of the format:

http://www.chemaxon.com/marvin/help/formats/cxsmiles-doc.html

 

1) Local bicyclo-alkane stereo information (local syn/anti, endo/exo representation). Please see the link above for more information. Yes, it is normal.

2) Yes, it recalculates all calculated fields based on the original structure stored in cd_structure.

3) It is the correct form, and most likely regeneration will produce the same cxsmiles

4) Usually an application developer should only use the following class when checking for regeneration and other upgrades at version change:

http://www.chemaxon.com/jchem/doc/dev/java/api/chemaxon/jchem/db/Updater.html

The regeneration process must produce exactly the same result all the time for correact search behaviour, so it cannot and should not be influenced.

NOTE:

The cd_smiles column in the JChem Base tables is for internal use only.

We do not recommend to use it by any application.

In certain cases it can also be NULL.

We recommend to use the original structure string in cd_structure for display and similar purposes.

 

Best regards,

 

Szilard