Descriptors Based on Three-Dimensional Structure

On the other hand, the Kohonen map in Fig. 19 allows only a limited view on similarity because pharmacological profiles of the drugs are not represented correctly. This is because pharmacological properties are not represented by the topological descriptor set. Descriptors based on three-dimensional structures and molecular interaction potentials (hydrogen bonds, lipophihc interactions, steric fit, etc.) are indispensable to describe these properties of molecules. 

For this reason, molecular interaction potentials based on three-dimensional structures are calculated for the compounds of the test dataset. The potentials include all possibibhes of interaction between the small molecule and the en-2ymes, as well as the shape of the molecule. Twenty autocorrelation coefficients are derived from these potentials for each molecule and used as a descriptor set. The new descriptors are mapped in the same way as the topological descriptors by means of a self-organizing map. 

This map (Fig. 21) displays a completely different picture. Now, the molecules are grouped by their pharmacological behavior instead of their structural scaffold. The COX-2-selective drugs (15, 16, 17, 19, 20) are grouped in the same region of the map. Compounds 14 (meloxicam) and 18 (6-mna) show a sHght selectivity for COX-2 and are located between the regions of selectivity and nonselectivity. 

---