Contents lists available at ScienceDirect Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio Robotic automation of a UHPLC/MS-MS method profling one-carbon metabolites, amino acids, and precursors in plasma Stephanie Andraos a , Michael Goy a , Benjamin B. Albert a , Martin Kussmann a,b , Eric B. Thorstensen a,∗ , Justin M. O'Sullivan a,b a Liggins Institute, The University of Auckland, New Zealand b New Zealand National Science Challenge “High-Value Nutrition”, New Zealand ABSTRACT Amino acids (AAs) and one-carbon (1-C) metabolism compounds are involved in a range of key metabolic pathways, and mediate numerous health and disease processes in the human body. Previous assays have quantifed a limited selection of these compounds and typically require extensive manual handling. Here, we describe the robotic automation of an analytical method for the simultaneous quantifcation of 37 1-C metabolites, amino acids, and precursors using reversed-phase ultra-high-pressure liquid chromatography coupled with tandem mass spectrometry (UHPLC/MS-MS). Compound extraction from human plasma was tested manually before being robotically automated. The fnal automated analytical panel was validated on human plasma samples. Our automated and multiplexed method holds promise for application to large cohort studies. 1. Introduction In the 1930s, Donald Van Slyke and Robert Dillon were some of the frst scientists to develop a method for the analysis of free amino acids using crude analytical methods [1]. Since then, the comprehensive analysis of low-molecular weight metabolites (i.e. metabolomics) has progressed thanks to advanced technology, in particular the coupling of chromatography and mass spectrometry, to enable their accurate quantitation in biological samples [2,3]. Several metabolomics methods have been developed for amino acids and/or 1-C metabolism compound quantifcation in diferent body fuids (e.g. plasma, cerebrospinal fuid, red blood cells) [4,5]. However, to our knowledge, none of these have been robotically automated. Robotic automation initially requires technical expertise to implement, but it provides several signifcant advantages once confgured. Key amongst these is the fact that robots are inherently more accurate than humans for the manual processing and pipetting of large numbers of samples. Thus, the robotic automa- tion simultaneously provides a faster and more reproducible sample handling. Amino acids and 1-C metabolites are central to physiological pro- cesses in the human body (e.g. nutritional metabolism, molecular, en- docrine and neurological functions) [6]. Due to their ubiquitous in- volvement in biological processes, quantifying the concentration of amino acids and 1-C compounds in human plasma can inform on the physiological status of an individual. Specifc metabolic profles have been linked to health and disease outcomes [7,8]. Estimating nutritional status using dietary reporting alone is unreliable [9]. Therefore, an objective quantitation of nutritional metabolites is of utmost importance, to strengthen our current understanding of human physiology and nutritional metabolism. Plasma is the non-cellular component of human blood. As such, plasma contains virtually all human proteins (e.g. blood clotting, binding proteins) representing those expressed throughout the tissues in addition to glucose, vitamins, and other nutrients. Blood collection from human subjects is relatively easy and non-invasive making it particularly appealing in both clinical and research settings [10]. In a healthy population, blood metabolites (e.g. amino acids) typically maintain homeostatic levels when in a fasted state, but may fuctuate substantially in the post-prandial state [11]. Therefore, physiological extrapolations need to be made cautiously when interpreting the results of plasma metabolomics studies, as these can refect states of home- ostasis (in the fasted state), a postprandial response (in the fed state), or metabolic exchanges between tissues. Notably, plasma-tissue correla- tions of metabolites are not always reliable [11]. For example, meta- bolites quantifed in plasma are typically not able to be assigned to cells or tissues of origin. However, an increasing level of biological under- standing of tissue-specifc pathways has allowed for the development of mathematical models, extrapolating organ-specifc metabolite levels based on their plasma measurements [12,13]. Here we present an automated method that combines a robotic extraction of plasma samples with reversed-phase ultra-high-pressure liquid chromatography coupled with tandem mass spectrometry https://doi.org/10.1016/j.ab.2019.113558 Received 29 October 2019; Received in revised form 19 December 2019; Accepted 20 December 2019 ∗ Corresponding author. E-mail address: e.thorstensen@auckland.ac.nz (E.B. Thorstensen). Analytical Biochemistry 592 (2020) 113558 Available online 03 January 2020 0003-2697/ © 2020 Elsevier Inc. All rights reserved. T