Descriptor Selection and Encoding

The atom pair, ap, and topological torsion, tt, descriptors are selected for illustrative purposes in the similarity searching context. 

A limitation of the ap and tt descriptors is the specificity of the atom typing, e.g., benzoic acid and phenyltetrazole would not be perceived as very similar, even though carboxylates and tetrazoles are both anions at physiological pH. 

A fuzzier atom type participating in these descriptors has been defined that is pharmacologically relevant - the physicochemical type at near-neutral pH [24], which is one of the following seven binding property classes 1 = cation 2 = anion 3 = neutral hydrogen-bond donor 4 = neutral H-bond acceptor  

5 = polar atom (atoms which are both donors and acceptors, e.g., hydroxy oxygen or either donor or acceptor via tautomerization, e.g., the nitrogens of imidazole)  

6 = hydrophobe 7 = other (nonpolar atoms in a polar environment or polar atoms that cannot accept or donate H-bonds). The physicochemical atom type, however, is too fuzzy for an atomic descriptor for the purpose of identifying common substructures (symmetric local similarity based on atoms and topological distances as descriptors). In this case, an atom type has been defined as a string containing the chemical element (all halogens equivalent to the element Hal), the number of incident r-bonds, and the physicochemical type. 

---