Alternative methods for estimating common descriptors for QSAR studies of dyes and fluorescent probes using molecular modeling software: 1. Concepts and procedures.

Quantitative structure activity relations (QSAR) models were developed to predict uptake and intracellular localization of probes or dyes in living cells. Many of the QSAR parameters used in such models are determined manually. Unfortunately, this requires a depth of chemical knowledge that biologists who wish to use these predictive tools do not necessarily possess. Moreover, some of the parameters are not easily obtained for all dyes and probes, which further restricts widespread use of QSAR methodology. Alternatives to some of these QSAR descriptors are defined and explained here. Estimation of these novel parameters using molecular modeling software, widely available and readily usable on personal computers in a variety of forms and brands, is described here. QSAR researchers need only draw the molecular structure and, with the proper commands, obtain either the parameters directly or the information to calculate them. I also demonstrate how the same software can generate some of the standard QSAR parameters, e.g., MW, Z, CBN, more reliably and conveniently than the manual procedures. A particularly problematic descriptor is log P, the logarithm of the octanol/water partition coefficient of a probe. This is discussed in detail and a novel alternative measure, the hydrophilic/lipophilic index (HLI), is introduced together with preliminary validation.