The attached protocol generates fingerprint features from a molecule.  I then take these features and try and map them back on to the original molecule to get feature counts.  I've tried both "substructure map" and "substructure count" and they give me the same odd result.  17 of the 55 features generated don't map back on to the molecule.  I've played about with the settings within the components and I can't get these to map.  Is this correct?  I find it strange that I can generate a fragment but then not map back on to it.  
​​​​​​​