Addressing drug-like/lead-like properties of biologically active small molecules early in a lead generation program is the current paradigm within the drug discovery community. Lipinski's "rule of five" has become the most commonly used tool to assess the relationship between structures and drug-like properties. Sixty percent of the 126 140 unique compounds in The Dictionary of Natural Products had no violations of Lipinski's "rule of five". We have isolated 814 natural products based on their expected drug-like/lead-like properties to generate a natural product library (NPL) in which 85% of the isolated compounds had no Lipinski violations. The library demonstrates the feasibility of obtaining natural products known for rich chemical diversity with the required physicochemical properties for drug discovery. The knowledge generated in creation of the library of structurally characterized pure natural products may provide opportunities to front-load lead-like property space in natural product drug discovery programs.

An NMR protocol that uses the residual proton signal from DMSO -d(6) (i.e., DMSO -d(5)) to determine the concentration of an analyte in a NMR sample was developed. This technique provides an alternative method for determining the molar concentration of compounds in solution without prior knowledge of their molecular weight. The method is particularly useful when submilligram quantities of compound are to be analyzed and is applicable to a variety of different research areas such as compound management, and natural product, combinatorial, and medicinal chemistry.

A robust method was developed to cluster similar NMR spectra from partially purified extracts obtained from a range of marine sponges and a plant biota. The NMR data were acquired using microtiter plate NMR (VAST) in protonated solvents. A sample data set which contained several clusters was used to optimize the protocol. The evaluation of the robustness was performed using three different clustering methods: tree clustering analysis, K-means clustering and multidimensional scaling. These methods were compared for consistency using the sample data set and the optimized methodology was applied to clustering of a set of spectra from partially purified biota extracts.