Skip to content

BCUT-descriptor gives different values depending on SMILES coding of same molecule #735

@lisakersten

Description

@lisakersten

Dear CDK-Developers,

I am having an issue regarding the BCUT-Descriptor.

Working with the BCUT-Descriptor showed that the BCUT-descriptor values depend on the SMILES coding. The same molecule (isomorphic structure) coded in different SMILES gives different BCUT-values:

Example: Acetamide

// use outgoing (old) smiles 
IAtomContainer tmpOutgoingAcetamide = sp.parseSmiles("C(C)(N)=O");
AtomContainerManipulator.convertImplicitToExplicitHydrogens(tmpOutgoingAcetamide );
     
// use absolute smiles generated using CDKs SmilesGenerator
IAtomContainer tmpAbsoluteAcetamide = sp.parseSmiles("CC(=O)N");
AtomContainerManipulator.convertImplicitToExplicitHydrogens(tmpAbsoluteAcetamide);
        
// calculate BCUT
BCUTDescriptor tmpBCutDesc = new BCUTDescriptor();
String resultBcutOutgoingAcetamide = tmpBCutDesc.calculate(tmpOutgoingAcetamide ).getValue().toString();
String resultBcutAbsoluteAcetamide = tmpBCutDesc.calculate(tmpAbsoluteAcetamide).getValue().toString();

gives:
for resultBcutOutgoingAcetamide : 
11.98891018734743, 16.005924145494618, -0.36108523358916594, 0.2733616424867599, 3.3749867502901925, 4.991266689763892
for resultBcutAbsoluteAcetamide : 
11.993832778266086, 15.994915651911281, -0.29718469746032156, 0.21853991312718152, 3.40274753607958, 4.981220894

more examples as .pdf and .csv can be found here:
Dependency_BCUT_SMILES.zip

Since the BCUT-descriptor is the only descriptor whose values depend on the SMILES coding, I wanted to ask whether it is necessary to generate absolute SMILES in advance or could it be an internal issue?

Best regards,
Lisa

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions