Analysis of a Large Structure-Activity Data Set Using Recursive Partitioning

Conventional parametric methods such as linear regression have not been entirely successful in analyzing structure‐activity data sets. This is because the underlying relationships may involve nonlinearities, thresholds and interactions, all of which considerably impede linear additive modelling approaches. Recursive partitioning, RP, is able to accommodate all these modelling difficulties seamlessly and therefore invites investigation as a general approach for study of structure activity relationships. In this paper we apply a recursive partitioning procedure, FIRM, to a large monoamine oxidase structure‐activity data set. The methodology is successful in uncovering nonlinearities in the response. Coupling RP with the use of correspondence analysis provides further insights into the distinction between compounds that are inactive, moderately active and active.