Classical QSAR methods describe structure-activity relationships in te
rms of physicochemical parameters and steric properties (Hansch analys
is, extrathermodynamic approach), or certain structural features (Free
Wilson analysis). 3D QSAR methods, especially comparative molecular f
ield analysis, consider the three-dimensional structures and the bindi
ng modes of protein ligands. Quantitative similarity-activity relation
ships derive correlations between the similarities of individual compo
unds and their biological activities. Theory and methodology of these
approaches are described here, together with the proper use of regress
ion and partial least squares analyses for deriving quantitative struc
ture-activity relationships. Part 2, to be published in the December i
ssue, will address applications and problems.