An Evaluation Of The Inter-Laboratory Reproducibility Of A Targeted Metabolomics Platform For Analysis Of Human Serum And Plasma

Hector KEUN, Imperial College, United Kingdom
SISKOS A. Department of Surgery & Cancer, Imperial College London , JAIN P. Department of Surgery & Cancer, Imperial College London

1 Department of Surgery & Cancer, Imperial College London

A critical question facing metabolomic research is whether data obtained from different centres can be effectively compared and combined, enhancing the statistical power and return­on­ investment of metabolomics studies. One important part of addressing this question is the assessment of the inter­laboratory precision (reproducibility) of the analytical protocols used.
Several test materials, including the NIST reference human plasma (SRM 1950), were distributed to six laboratories and independently analysed using the AbsoluteIDQTM p180 Kit (Biocrates Life Sciences AG). This LC-MS/MS platform allows the targeted analysis of amino acids, biogenic amines, acylcarnitines, sphingolipids and glycerophospholipids.

After excluding 12 metabolites (of 189) not consistently detected in all laboratories, a high degree of analytical precision was observed across for metabolites measured quantitatively. Normalisation of measurements to the profile of a standard reference material obtained in each laboratory run significantly improved the inter­laboratory precision of the lipid profile measured. After normalisation, the majority (typically ~75%) of metabolites in each test material exhibited an inter­laboratory coefficient of variance (CV) of <10%. Approximately 90% of metabolites exhibited an inter­laboratory coefficient of variance (CV) of <20%. Ongoing analysis will also assess the impact of highly lipidic samples and varying anti­coagulant on precision, as well as the accuracy of measurement of specific quantified metabolites. This is the first inter­laboratory assessment of this metabolomics platform, providing critical information for users to interpret these data appropriately.