This Analysis compares two large-scale pharmacogenomic data sets that catalogued the sensitivity of a large number of cancer cell lines to approved and potential drugs, and finds that whereas the gene expression data are largely concordant between the two studies, the reported drug sensitivity measures and subsequently their association with genomic features are highly discordant. Two large-scale data sets recently catalogued the sensitivity of a large number of cancer cell lines to pharmacological drugs, and integrated the drug-response data with genomic features, such as mutations and gene expression profiles. This Analysis by John Quackenbush and colleagues compares the two studies and finds that although the gene expression data are largely concordant between them, the reported drug-sensitivity measures and subsequently their association with genomic features are highly discordant. The authors call for standardization of drug-response measurements, to aid the discovery of robust biomarkers and mechanisms of drug response and hence progress in personalized cancer medicine. Two large-scale pharmacogenomic studies were published recently in this journal. Genomic data are well correlated between studies; however, the measured drug response data are highly discordant. Although the source of inconsistencies remains uncertain, it has potential implications for using these outcome measures to assess gene–drug associations or select potential anticancer drugs on the basis of their reported results.

Inconsistency in large pharmacogenomic studies