FINDING OPTIMUM FIELD MODELS FOR 3D QSAR

A field optimization paradigm for 3D QSAR is described. This study inv oked nine 3D fields: steric, electrostatic, HINT (hydropathic), E-Stat e, HE-State, H-Bond (field couplet), and indicator (field couplet); in an examination of ryanoids. Several combinations of fields yield impr oved q(2) (cross-validated r(2)) over the standard steric/electrostati c CoMFA treatment with improvements of 6-10% in q(2) being typical. Th e choice of field model is shown ultimately to be a function of the de sired model characteristics as many of these fields have overlapping i nformation.