Structure-controlled automated purification of parallel synthesis products in drug discovery

Extensive automation of both random and rational drug discovery strategies greatly increases the number of compounds entering biological screens. Although parallel synthesis (one compound per well) strategies eliminate the deconvolution step necessary when pooled libraries are screened, parallel synthesis products are usually screened as crude mixtures, because purification slows the process of lead discovery. Screening crude products is sometimes complicated by synergies and interferences between compounds. Screening pure compounds is the only sure route to immediate, reliable structure - activity relationships. Automated purification strategies are designed to limit or remove the purification bottleneck between synthesis and screening. In this paper, a workstation is described which uses a combination of UV absorbance and mass spectrometric data to make real-time decisions for HPLC fraction collection, allowing selection of compounds based on mass or substructure. This methodology has demonstrated success with parallel synthesis products in drug discovery applications. (C) 1998 John Wiley & Sons, Ltd.