The process of metabolite identification is essential to the drug discovery and development process; this is usually achieved by liquid chromatography/tandem mass spectrometry (LC/MS/MS) or a combination of liquid chromatography/mass spectrometry (LC/MS) and nuclear magnetic resonance (NMR) spectroscopy. Metabolite identification is, however, a time-consuming process requiring an experienced skilled scientist. Multivariate statistical analysis has been used in the field of metabonomics to elucidate differences in endogenous biological profiling due to a toxic effect or a disease state. In this paper we show how a combination of liquid chromatography/time-offlight mass spectrometry (LC/TOFMS) and multivariate statistical analysis can be used to detect drug metabolites in a biological fluid with no prior knowledge of the compound administered. Copyright # 2003 John Wiley & Sons, Ltd.As part of the drug discovery and development process it is essential to identify the pharmacokinetic and metabolic characteristics of the candidate pharmaceutical for selection and subsequent regulatory submissions. Failure to identify a drug metabolite can have very serious implications; for instance, the metabolite may actually be the active pharmacophore in the systemic system, or it could be a marker of toxicity. Therefore, the detection, identification and profiling of drug metabolites has become a central part of the drug discovery and development process.The process of metabolite detection and identification is typically a labor-intensive and time-consuming process. This process has been simplified by the use of radiolabeled compounds 1 and/or spectroscopic techniques such as mass spectrometry 2,3 and NMR spectroscopy. 4-6 These approaches have allowed drug metabolism scientists to detect/identify compounds at lower concentrations and with greater speed. Of these analytical techniques, LC/MS and LC/MS/MS have become the most popular and widely employed due to their speed, sensitivity and specificity. However, the use of these techniques for metabolite detection and identification still requires a trained and skilled scientist to produce highquality results.With the advent of combinatorial chemistry the number of compounds entering the drug discovery process for candidate evaluation has significantly increased. This has significantly increased the number of compounds submitted for metabolite profiling and identification. In an attempt to simplify and streamline this process, many analytical instrument manufacturers have developed software packages that allow the automated review of LC/MS data such that potential drug metabolites are identified.7 These software packages typically accomplish drug metabolite identification by calculating molecular masses of postulated drug metabolites through the addition of classical metabolic transitions to the mass of the administered compound, followed by the subsequent extraction of these masses from the TIC LC/MS data set, e.g., parent mass þ16 Da for hydroxylation, þ80 Da for sulfation, þ176...