Q-MODEL CLUSTER ANALYSIS OF 96 COMMON AROMATIC SUBSTITUENTS

In drug design, the large number of possible substituents that might be selected for an initial set of derivatives presents a formidable problem in decision making. By factoring such a set into more or less homogeneous subgroups with respect to various chemical structural parameters of importance, one can focus upon such special considerations as hydrophobic effects, metabolic behavior, or ease of synthesis. If the clusters are formed by an objective procedure such as resemblance coefficient between the points in a parameter space, selecting one derivative from each cluster will tend to give a maximum range in parameter type and help to establish a viable structure-activity relationship more rapidly. In this paper, 96 common aromatic substituents have been clustered into 5,10, and 18 clusters with respect to various combinations of length L (A) in Sterimol system, the lipophilie π constant, electronic Swain and Luptontype F and R constants, σ_m and σ_p constants. The dendrogram of classification is also provided. These clusters may be applied to selecting molecules in drug design.