Scheubert, K. et al. Nat. Commun. 8, 1494 (2017).

In both untargeted mass-spectrometry-based proteomics and metabolomics studies, experimental tandem mass spectra are matched to reference library spectra for interpretation. Manual checking of these matches becomes impractical in large-scale studies; as a result, statistical methods for estimating the false discovery rate (FDR) of this matching process have become essential for proper reporting of proteomics results. However, equivalent approaches for assessing FDR in metabolomics studies have lagged far behind. Scheubert et al. now report FDR estimation methods for metabolomics based on empirical Bayes statistics and on the target–decoy approach (which is widely applied in proteomics for FDR estimation). They applied these approaches to 70 publicly available metabolomics data sets and found that scoring parameters must be adjusted for each data set. Statistical validation will help advance the reliability and robustness of metabolomics results.