Translational biomarker discovery in clinical metabolomics: an introductory tutorial

• DOI: 10.1007/s11306-012-0482-9
• Publication Date: 2012-12-03T16:39:48Z
• Authors: Jianguo Xia; David I. Broadhurst; Michael Wilson; David S. Wishart
• Keywords: Biomarker Discovery

Metabolomics is increasingly being applied towards the identification of biomarkers for disease diagnosis, prognosis and risk prediction. Unfortunately among many published metabolomic studies focusing on biomarker discovery, there very little consistency relatively rigor in how researchers select, assess or report their candidate biomarkers. In particular, few any measure sensitivity, specificity, provide receiver operator characteristic (ROC) curves with associated confidence intervals. Even fewer explicitly describe release model used to generate ROC curves. This surprising given that most other biomedical fields, curve analysis generally considered standard method performance assessment. Because ultimate goal discovery translation those clinical practice, it clear metabolomics community needs start "speaking same language" terms reporting-especially if wants see metabolite markers routinely clinic. this tutorial, we will first introduce concept use single chemistry. includes construction curves, understanding meaning area under (AUC) partial AUC, as well calculation The second part tutorial focuses analyses within context metabolomics. section describes different statistical machine learning strategies can be create