Structural dissimilarity index

Snb-stmctnre diversity is most easily defined using metrics such as the Tanimoto Dissimilarity Index. These metrics are based on linear bitmaps (fingerprints) generated from the molecnlar fragments or compound sub-structures (Figure 3). This approach has been developed extensively by Daylight Chemical Information Systems. ... 

Ob-6 AvatyCO 1) In the QSAR-context, it is necessary and sufficient that p verifies conditions C2), i.e., to be a dissimilarity index. In this sense MSD was used as a steric parameter (instead of the Taft Eg constant, for example) in Hansch type structure-biological activity correlations. 2) These results obtained for MSD are valid... 

Aquatic microcosms and mesocosms offer the ideal situation to investigate populations of species interacting in their natural environment (i.e., to study communities stressed in structured systems see Section 4.5.1). It is, however, only recently that these experiments have been analyzed at the community level (Van Wijngaarden et al. 1995 Sparks et al. 1999 van den Brink and Ter Braak 1999). Until 10 years ago, experiments were evaluated at the population level, largely ignoring species interactions and energy flows in the systems. The development of community-level endpoints offered the possibility to evaluate the experiments on a community level (i.e., they offered the opportunity to scale up the level of evaluation Kedwards et al. 1999). Summary community-level endpoints calculated from the results of these experiments are mostly structural ones measures of diversity (e.g. numbers of species, and the Shannon-Weaver diversity index) and similarity of the treated systems compared to the untreated controls (e.g., the principal response curves method, Bray-Curtis dissimilarity, or Stander s index see van den Brink and Ter Braak 1998 for a comparison). 

In recent years there has been a notable increase in research on structure-activity relationships (SARs), also called quantitative structure-activity relationships (QSARs), used to assess the toxicity of substances for which there are few experimental data. This approach involves establishing mathematical relationships derived from computer modeling, based on known toxicity data of similar (or dissimilar) types of compounds, octanol-water partition coefficients, molar connectivity index values, and other parameters. A detailed discussion on this subject is beyond the scope of this book. 

Optical properties also provide usefiil structure information about chemically similar fibers. The orientation of the molecular chains of a fiber can be estimated from differences in the refractive indexes measured with the optical microscope, using hght polarized in the parallel and perpendicular directions relative to the fiber axis (39,40). The difference of the principal refractive indexes is called the birefringence, which is generally used to monitor the orientation of nylon filament in melt spinning (43). Birefingence can also be used to distinguish nylon from other chemically dissimilar fiber types as illustrated ... 

The second method to be discussed, MSA, was introduced in 1980. It explicitly assumes that the shapes of the molecules provide information about the shape of the receptor cavity.Pairwise similarities and dissimilarities are calculated between a reference structure and the other compounds of the data set. Typically, biological potency is correlated with one shape index plus conventional QSAR descriptors for lipophilicity and electronic effects. 

Much research has been done into similarity searching, and it is often assumed that diversity, or dissimilarity, is the converse (i.e, 1 — similarity). Any similarity measure involves three main components the structural descriptors that are used to characterize the molecules, the weighting scheme used to differentiate important from less important characteristics, and the similarity coefficient that is used to quantify the degree of similarity between pairs of molecules. Three types of descriptor have been used fragment substructures, topological indexes, and global physical properties. One of the most common similarity coefficients is that due to Tanimoto. There are three main methods of compound selection cluster-based, dissimilarity-based, and partition-based. 

---