Inter-laboratory reproducibility of a targeted metabolomics platform for analysis of human serum and plasma

Abstract

A critical question facing the field of metabolomics is whether data obtained from different centres can be effectively compared and combined. An important aspect of this is the inter-laboratory precision (reproducibility) of the analytical protocols used. We analysed human samples in six laboratories using different instrumentation but a common protocol (the Absolute IDQ™ p180 Kit) for the measurement of 189 metabolites via liquid chromatography (LC) or flow-injection analysis (FIA) coupled to tandem mass spectrometry (MS/MS). In spiked quality control (QC) samples 82% metabolite measurements had an inter-laboratory precision of <20%, while 83% of averaged individual laboratory measurements were accurate to within 20%. For 20 typical biological samples (serum and plasma from healthy individuals) the median inter-laboratory CV was 7.6%, with 85% of metabolites exhibiting a median inter-laboratory CV of <20%. Precision was largely independent of the type of sample (serum or plasma) or the anticoagulant used but was reduced in a sample from a patient with dyslipidaemia. The median inter-laboratory accuracy and precision of the assay for standard reference plasma (NIST SRM 1950) were 107% and 6.7%, respectively. Likely sources of irreproducibility were the near-LOD typical abundance of some metabolites and the degree of manual review and optimisation of peak integration in the LC-MS/MS data post-acquisition. Normalisation to a reference material was crucial for the semi-quantitative FIA measurements. This is the first inter-laboratory assessment of a widely-used, targeted metabolomics assay illustrating the reproducibility of the protocol and how data generated on different instruments could be directly integrated in large-scale epidemiological studies.