The much publicized "Rule of 5" has been widely adopted among the pharmaceutical industry. It is used as a first step filter to perform virtual screening of compound libraries, in an effort to quickly eliminate lead candidates that have poor physicochemical properties for oral bioavailabilty. One of the key parameters used therein is log P, which is a useful descriptor, but one that fails to take into account variation in the lipophilicity of a drug with respect to the ionic states present at key biological pH values. Given that the majority of commercial pharmaceuticals contain an ionizable moiety, we propose that log D is a better descriptor for lipophilicity in the context of the Rule of 5. It gives more physiologically relevant results, thereby reducing the number of potential false-negatives incorrectly eliminated in screening. Using a series of commercial compound libraries, this study showed that the adapted Rule of 5 using log D instead of log P provides notable improvement in pass rate for compounds that have the desired lipophilicity at a relevant physiological pH.

A computational study was undertaken to discern where and how chiral alkanes, alcohols, and acetates enantioselectively bind to permethyated -cyclodextrin, the most commonly used chiral stationary phase in gas chromatography. We found that enantioselective binding data could be reproduced with standard molecular dynamics techniques if averages are taken over multiple trajectories of nanosecond simulation times each, while Metropolis Monte Carlo simulations using rigid body molecules are unable to reproduce chromatographic retention orders. Data extracted from the molecular simulations revealed the preferred binding site for small analytes to be the interior of the macrocycle, with rapid shuttling between the primary and secondary rims and low-energy excursions into and out of the host cavity. The dominant forces holding the host-guest complexes together are the short range dispersion forces. The enantiodiscriminating forces responsible for chiral recognition are also the short range van der Waals forces and these enantiodifferentiating forces are typically 1-2 orders of magnitude smaller than the binding forces. An assessment of the number of hydrogen bonds for the diastereomeric complexes is presented along with the locations of dominant hydrogen-bonding sites on the macrocycle. A comparison is made between analytes capable of intramolecular hydrogen bonding with those that can not. It is pointed out that the 3-point binding description of chiral discrimination can be used, but it loses its appeal at such high temperatures due to ill-defined structures.

We report several binary classification models that directly link the genetic toxicity of a series of 140 thiophene derivatives with information derived from the compounds' molecular structure. Genetic toxicity was measured using an SOS Chromotest. IMAX (maximal SOS induction factor) values were recorded for each of the 140 compounds both in the presence and in the absence of S9 rat liver homogenate. Compounds were classified as genotoxic if IMAX >or= 1.5 in either test or nongenotoxic if IMAX < 1.5 for both tests. The molecular structures were represented by numerical descriptors that encoded the topological, geometric, electronic, and polar surface area properties of the thiophene derivatives. The classification models used were linear discriminant analysis (LDA), k-nearest neighbor classification (k-NN), and the probabilistic neural network (PNN). These were used in conjunction with either a genetic algorithm or a generalized simulated annealing to find optimal subsets of descriptors for each classifier. The quality of the resulting models was determined by the number of misclassified compounds, with preference given to models that produced fewer false negative classifications. Model sizes ranged from seven descriptors for LDA to three descriptors for k-NN and PNN. Very good classification results were obtained with all three classifiers. Classification rates for the LDA, k-NN, and PNN models were 80, 85, and 85%, respectively, for the prediction set compounds. Additionally, a consensus model was generated that incorporated all three of the basic model types. This consensus model correctly predicted the genotoxicity of 95% of the prediction set compounds.

A method for molecular dynamics (MD), Monte Carlo (MC), and energy minimization simulation utilizing a Hamiltonian that is divided into two parts is described. One part is treated with a quantum mechanical Hamiltonian, typically a small part of the simulated system that comprises the chromophore. The other part is treated with a classical mechanical Hamiltonian. This partitioning of the system allows us to simulate, for example, not only electronic spectroscopy but also chemical reactions where a bond is broken or to explore the excited state potential energy surface. The particular choice of the quantum mechanical Hamiltonian, the intermediate neglecting of differential overlap (INDO) model Hamiltonian, also offers the possibility of simulating systems that contain a transition metal, which only rarely have been accessible with traditional MD and MC methods. Test calculations on small systems are presented together with an investigation of the photophysics of uracil and 1,3-dimethyluracil.