Properties fragment-based

A particularly good selection of physical properties may be spectra, because they are known to depend strongly on the chemical structure. In fact, different types of spectra carry different kinds of structural information, NMR spectra characterize individual carbon atoms in their molecular environment. They therefore correspond quite closely to fragment-based descriptors, as underlined by the success of approaches to predict NMR spectra by fragment codes (see Section 10.2.3). 

Ertl P, Rohde B, Selzer P. Fast calculation of molecular polar surface area as a sum of fragment-based contributions and its application to the prediction of drug transport properties. I Med Chem 2000 43 3714-7. 

In addition to physical properties, substructure-based filters can be applied to reduce further the number of molecules, for instance molecules with undesirable functionality for example, reactive or toxic groups can be removed and molecules with particular features (or atoms) can be actively selected. There may be particular functionality that it is desirable to avoid due to assay format, such as fluorophores in fluorescence-based approaches. Structural features for these inclusion and exclusion criteria can be readily formulated using SMILES-based procedures and this type of substructure-based compound selection technique can also be employed in the generation of focused sets of fragment molecules. 

Many different structural descriptors have been developed for similarity searching in chemical databases [4] including 2D fragment based descriptors, 3D descriptors, and descriptors that are based on the physical properties of molecules. More recently, attention has focused on diversity studies and many of the descriptors applied in similarity searching are now being applied in diversity studies. Structural descriptors are basically numerical representations of structures that allow pairwise (dis)similarities between structures to be measured through the use of similarity coefficients. Many diversity metrics have been devised that are based on calculating structural (dis)similarities, some of these are described below. 

---