Three major dynamics are reshaping the pharmaceutical industry today. First, the Human Genome Project is making novel protein targets available on a grand scale. Second, combinatorial chemistry is allowing access to molecular diversity on an even grander scale--a scale larger than can be handled via high-throughput screening. And third, pharmaceutical scientists are increasingly recognizing the importance of patient-to-patient differences (genetic polymorphisms) in determining the competitive differentiation among competing drugs directed at the same pharmacological target. SBI's technology makes it possible for pharmaceutical companies to harness these three dynamics by providing unique technologies that unify genomics, combinatorial chemistry, and screening to accelerate lead finding and optimization through the use of protein structure.

For drug designers, protein chemists, pharmacologists, pharmacogenomicists, molecular biologists--virtually all scientists in the life science field--an understanding of three-dimensional protein structure is the key to understanding, and productively manipulating, biological function. This includes, for example, designing and optimizing drugs; guiding site-directed mutagenesis and biochemical labeling; and predicting cross-reactivity (side-effects), pharmacogenomic differences, immunogenicity, stability and other physical properties of proteins.

SBI has developed a proprietary capability to accurately model protein surface structural information on an industrial scale using proprietary ab initio loop prediction and local tertiary structure prediction algorithms. In addition, the Company's DynaPharm^TM technology permits the extraction of drug structure information from dynamic protein surface information derived from proprietary SBI models, or from X-ray crystallographic or NMR structures.

Use of the DynaPharm^TM technology in combination with SBI's high-quality surface structural information has led repeatedly, over the last eighteen months, to the generation of small-molecule drug leads from gene sequences for a variety of drug targets. In each case, this has been done in a matter of months and with a 100- to 1000-fold reduction in the number of compounds necessary to synthesize and screen as compared with current pharma industry practice. DynaPharm^TM technology permits the generation of active molecules built around multiple chemical scaffolds. The use of DynaPharms to screen computationally various virtual combinatorial libraries (CombiLib^TM modules) built around a large number of chemical scaffolds considered attractive for pharmaceutical application allows SBI to unlock the full potential of molecular diversity available through combinatorial chemistry. Because computational screening of existing compound collections or SBI's virtual CombiLib^TM modules typically yields hit rates in actual laboratory screening of 5%-20%, only a small number of molecules (~50-100) need be synthesized or selected from existing compound collections, resulting in perhaps 10 to 20 active hits. By comparison, the typical hit rate in high-throughput screening (HTS) is 0.1% to 0.01% (1 per thousand to 1 per 10,000 molecules screened). Thus, SBI effects savings in actual testing and compound acquisition or synthesis costs of 99.9% compared to HTS. The need to synthesize only a small number of molecules to generate a large number of hits means that SBI can address new and multiple chemical scaffolds without the need to synthesize tens of thousands of molecules to sample each new scaffold of interest. This translates to speedy and economical generation of broader offensive and defensive patent coverage and a greater number of backup chemical series for each target. SBI's structure-informed approach to drug discovery and optimization has resulted in the repeated and reliable generation of multiple classes of small-molecule antagonists for a wide range of targets, including enzyme, protein-protein interaction and hormone targets.

SBI launched its first database product, the SBdBase^TM, in the first half of 1999. The SBdBase^TM is intended to provide a continually expanding database of high-quality protein structural information and is intended to be comprehensive for the human genome and related proteins within a reasonable timeframe. SBI uses comparative modeling substantially augmented with proprietary energetics and molecular mechanics-based ab initio loop computations and local tertiary structure prediction to refine SBdBase models for immediate use in structure-based drug design.

An increasing percentage of protein structures are accessible for research use through comparative modeling. SBI is a leader in development of new and increasingly powerful methods for extending the range of computationally accessible protein structures. The computational methods at SBI yield many energetically allowed conformations in a dynamic trajectory, including those found by static physical methods like X-ray crystallography. The SBdBase^TM currently contains thousands of proprietary high-quality protein structures built around more than 175 protein families selected to be of greatest interest to the drug discovery field. Structures and families are added on a continuing basis. For those protein families included in the SBdBase^TM, the models are shown to be comparable to crystal or NMR structures. They have been used repeatedly and reliably in the generation of novel drug leads in a wide range of target classes including, for example, the growth factor, viral protease, protein kinase, hormone and signal transduction areas.