Characterization of missing values in untargeted MS-based metabolomics data and evaluation of missing data handling strategies
BACKGROUNDUntargeted mass spectrometry (MS)-based metabolomics data often contain missing values that reduce statistical power and can introduce bias in epidemiological studies. However, a systematic assessment of the various sources of missing values and strategies to handle these data has received little attention. Missing data can occur systematically, e.g. from run day-dependent effects due to limits of detection (LOD); or it can be random as, for instance, a consequence of sample preparation.\n\nMETHODSWe investigated patterns of missing data in an MS-based metabolomics experiment of serum samples from the German KORA F4 cohort (n = 1750). We then evaluated 31 imputation methods in a simulation framework and biologically validated the results by applying all imputation approaches to real metabolomics data. We examined the ability of each method to reconstruct biochemical pathways from data-driven correlation networks, and the ability of the method to increase statistical power while preserving the strength of established genetically metabolic quantitative trait loci.\n\nRESULTSRun day-dependent LOD-based missing data accounts for most missing values in the metabolomics dataset. Although multiple imputation by chained equations (MICE) performed well in many scenarios, it is computationally and statistically challenging. K-nearest neighbors (KNN) imputation on observations with variable pre-selection showed robust performance across all evaluation schemes and is computationally more tractable.\n\nCONCLUSIONMissing data in untargeted MS-based metabolomics data occur for various reasons. Based on our results, we recommend that KNN-based imputation is performed on observations with variable pre-selection since it showed robust results in all evaluation schemes.\n\nKey messagesO_LIUntargeted MS-based metabolomics data show missing values due to both batch-specific LOD-based and non-LOD-based effects.\nC_LIO_LIStatistical evaluation of multiple imputation methods was conducted on both simulated and real datasets.\nC_LIO_LIBiological evaluation on real data assessed the ability of imputation methods to preserve statistical inference of biochemical pathways and correctly estimate effects of genetic variants on metabolite levels.\nC_LIO_LIKNN-based imputation on observations with variable pre-selection and K = 10 showed robust performance for all data scenarios across all evaluation schemes.\nC_LI