New developments in a hazard identification algorithm for hormone receptorligands

Recently we described the Common REactivity PAttern (COREPA) technique to s creen data sets of diverse structures for their ability to serve as ligands for steroid hormone receptors [1]. The approach identifies and quantifies similar global and local stereoelectronic characteristics associated with a ctive ligands through a comparison of energetically-reasonable conformer di stributions for selected descriptors. For each stereoelectronic descriptor selected, discrete conformer distributions from a training set of ligands a re evaluated and parameter ranges common for conformers from all the chemic als in the training set are identified. The use of discrete partitions of p arameter ranges to define common reactivity patterns can, however, influenc e the outcome of the algorithm. To address this limitation, the original me thod has been extended by approximating continuous conformer distributions as probability distributions. The COREPA-Continuous (COREPA-C) algorithm as sesses the common reactivity pattern of biologically similar molecules in t erms of a product of probability distributions, rather than a collection of common population ranges determined by examination of discrete partitions of a distribution. To illustrate the algorithm, common reactivity patterns based on interatomic distance and charge on heteroatoms were developed and evaluated using a set of 28 androgen receptor ligands. Notable attributes o f the COREPA-C algorithm include flexibility in establishing stereoelectron ic descriptor criteria for identifying active and nonactive compounds and t he ability to quantify three-dimensional chemical similarity without the ne ed to predetermine a toxicophore or align compounds(s) to a lead ligand.