Correct number of stereoisomers with InChiKey canonizer

Hi,

using Marvin 5.2.4 the following mol (a cardiolipin) calculates 16 stereoisomers,

under Marvin 5.1.0 it was 10 stereoisomers, however the correct number is 3 (I guess)

SMILES:

[H]C(O)(COP(O)(=O)OCC([H])(COC(=O)CCC)OC(=O)CCC)COP(O)(=O)OCC([H])(COC(=O)CCC)OC(=O)CCC

 

Its highly symmetric and the following names are generated with MView 5.1.0

[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid
[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid

[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2S)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid
[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2S)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid

[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid
[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid

[(2S)-2,3-bis(butanoyloxy)propoxy][3-({[(2S)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid

[(2R)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid
[(2R)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid
[(2R)-2,3-bis(butanoyloxy)propoxy][3-({[(2R)-2,3-bis(butanoyloxy)propoxy](hydroxy)phosphoryl}oxy)-2-hydroxypropoxy]phosphinic acid

Or InChiKey generated with OpenBabel due to missing InchiKey support in Marvin :-(

KGHYARREBKRHGB-OISBFRGYDW
KGHYARREBKRHGB-OISBFRGYDW

KGHYARREBKRHGB-VALLNNISDM
KGHYARREBKRHGB-VALLNNISDM

KGHYARREBKRHGB-OISBFRGYDW
KGHYARREBKRHGB-OISBFRGYDW

KGHYARREBKRHGB-VALLNNISDM

KGHYARREBKRHGB-YGKSKUPNDF
KGHYARREBKRHGB-YGKSKUPNDF
KGHYARREBKRHGB-YGKSKUPNDF

 

So after dupilcate removal using either the name or better the InChiKey one can get the three correct results. In case of the 4:1 isomers (one double bond) Marvin 5.2.4 creates 256 stereoisomers, and after filtering with the InChiKey only 64 stereoisomers are left.

SMILES for stereo enumeration example:

[H]C(O)(COP(O)(=O)OCC([H])(COC(=O)\C=C\C)OC(=O)\C=C\C)COP(O)(=O)OCC([H])(COC(=O)\C=C\C)OC(=O)\C=C\C

Tobias

 

 

Hi Tobias,

It seems unique smiles generation (unique smiles is used for duplicate filtering) does not work always correctly on highly symmetric molecules. We will check this issue.

Thanks for the report,

Zsolt

Hi Tobias,

TobiasKind wrote:

Or InChiKey generated with OpenBabel due to missing InchiKey support in Marvin :-(

InChIKey is supported in Marvin. In the version you mentioned (5.2.4) it is only reachable as an export option if inchi ("inchi:key"), but in the latest version 5.2.5 we implemented inchikey as a new export format ("inchikey") too.

Best Regards, Peter

Thats nice! I have been desperately waiting for the InChiKey DLL linkage.

Thank you

Tobias

 

Hi Tobias,

I think the main difference that we treat R1-O-P(=O)(O)-O-R2 group as a chiral center while others are not.  On the other hand, I think if we don't treat the phosphate esther group as chiral center we should have 4 isomers.

Andras