A primary goal of lead optimization is to identify compounds with improved absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties. A number of reports have linked computed molecular properties to desirable in vivo ADMET outcomes, but a significant limitation of these analyses is the failure to control statistically for possible covariates. We examine the relationship between molecular properties and in vitro surrogate assays vs in vivo properties within 173 chemical series from a database of 3773 compounds with rodent pharmacokinetic and toxicology data. This approach identifies the following pairs of surrogates as most predictive among those examined: rat primary hepatocyte (RPH) cytolethality/volume of distribution (V(d)) for in vivo toxicology outcomes, scaled microsome metabolism/calculated logP for in vivo unbound clearance, and calculated logD/kinetic aqueous solubility for thermodynamic solubility. The impact of common functional group substitutions is examined and provides insights for compound design.