Many post genomic clinical studies including metabolomics, transcriptomics, proteomics, and other high-content or high-throughput, 'omic experiments are set up such that the primary aim is the discovery of biomarkers that can discriminate, with a certain level of certainty, between nominally matched 'case' and 'control' samples. However, it is unfortunately very easy to find markers that are apparently persuasive but that are in fact entirely spurious. Easily avoidable bad practice in experiment design and execution can occur at many levels prior to statistical modeling, including:  bias in patient selection, inconsistent bio-banking, poor choice in clinical endpoint, lack of understanding of the limitations analytical platforms, inadequate sample size, inappropriate choice of statistical methods, inadequate model assessment, and minimal and ineffective lab based quality assurance protocols. Many studies fail to take these key factors into account, and thereby fail to discover anything of true significance, or more seriously, report spurious findings that prove impossible to validate. Here, I summarize these problems from a practical perspective, and provide pointers to assist in the improved design and evaluation of biomarker discovery experiments, with the emphasis on clinical metabolomics studies.

Better data by design: The critical role of experiment design, quality assurance, and model assessment in effective clinical metabolomics biomarker discovery

248th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.