This paper gives an overview of computer assisted molecule design using quantitative structure activity relationships (QSAR). In deriving a QSAR for a set of compounds, one should be aware of possible properties one can use to correlate with the activity of this set of compounds. The paper describes some of the traditional and also more recently used properties that can be calculated. The traditionally used properties have been those calculated for the substituent groups that were varied on a particular scaffold, whereas the more recently used properties are those calculated on the molecule as a whole. One of the most recently used properties for comparative purposes are fields around the molecules. The technique is termed comparative molecular field analysis (CoMFA). The statistical techniques used to correlate the properties with the activity are also described and, finally, a case study is described. The case study shows the application of traditional and 3D QSAR (CoMFA) techniques on some photosystem II inhibitors.

Although there are many glutamate receptors in the retina, 2-amino-4-phosphonobutyrate (L-AP4) is an agonist that acts selectively at metabotropic glutamate receptors (mGluR6) of ON bipolar cells. We explored the properties of agonists that activate this receptor. The effects of various glutamate analogs on the b-wave of the electroretinogram (ERG) were used as a measure of their activity. Conformational comparisons among agonists suggest that ligands in an extended conformation preferentially bind to the ON bipolar synaptic receptor. But this property is insufficient to explain the selectivity of mGluR6 because some inactive glutamate analogs could also match this extended conformation. Comparative molecular field analysis (CoMFA) was used to compare the electrostatic and steric potentials of agonists with their action at the ON bipolar synapse. Steric potentials beneath a plane defined by the three putative binding sites plays a key role in determining agonist activity. The CoMFA model was used to predict the activity of glutamate analogs and correlations between predicted and measured activity support the model.