October 21, 1998
Statistical Engineering Division
Information-intensive drug discovery: Genomics, proteomics, and bioinformatics

The current revolution in drug discovery is based largely on developments in molecular biology and informatics. A case in point is provided by the drug discovery program of the NCI, which has profiled more than 65,000 compounds for their activity against 60 human cancer cell lines. The resulting patterns of activity (pharmacological fingerprints) have proved rich in information on mechanisms of drug action and resistance (Paull, et al., JNCI 81:1088,1989; Weinstein, et al., Science 258:447, 1992). To characterize the 60 cell types (and selected transfectants) with respect to molecular markers, we have done careful parallel harvests of DNA, RNA, and protein for what I have termed "omic" analysis -- including protein expression profiling by 2-D gels and mRNA expression profiling by high density cDNA and oligonucleotide microarrays. The data complement those of the NCI's Cancer Genome Anatomy Project in that the 60 cell line "patients," unlike most human ones, have extensive, well-defined treatment histories -- i.e., they have been treated with >65,000 agents one at a time and independently (Weinstein, et al., Science 275:343, 1997).