Fig. 8
From: An algorithm to classify homologous series within compound datasets
The impact of different fragmentation approaches in the RDKit on homologous series core detection, top: using DeleteSubstructs, bottom: using ReplaceCore. The two input molecules are homologous and should be classified into the same series; fragmentation using ReplaceCore achieves this as identical cores are detected (C.C.O in SMILES representation). However, DeleteSubstructs yields different cores (C.C.O and C.C.CO in SMILES respectively) for the two input molecules because both –CH2–CH2–CH2–CH2– chains of the symmetrical molecule are removed simultaneously in Fragmentation Step 1, resulting in inequivalent cores and no homologous series detected