Christina Ranninger, Lukas E Schmidt, Marc Rurik, Alice Limonciel, Paul Jennings, Oliver Kohlbacher, and Christian G Huber (2016)
Improving global feature detectabilities through scan range splitting for untargeted metabolomics by high-performance liquid chromatography-Orbitrap mass spectrometry
Anal. Chim. Acta, 930:13-22.
Untargeted metabolomics aims at obtaining quantitative information on the highest possible number of low-molecular biomolecules present in a biological sample. Rather small changes in mass spectrometric spectrum acquisition parameters may have a significant influence on the detectabilities of metabolites in untargeted global-scale studies by means of high-performance liquid chromatography-mass spectrometry (HPLC-MS). Employing whole cell lysates of human renal proximal tubule cells, we present a systematic global-scale study of the influence of mass spectrometric scan parameters and post-acquisition data treatment on the number and intensity of metabolites detectable in whole cell lysates. Ion transmission and ion collection efficiencies in an Orbitrap-based mass spectrometer basically depend on the m/z range scanned, which, ideally, requires different instrument settings for the respective mass ranges investigated. Therefore, we split a full scan range of m/z 50-1000 relevant for metabolites into two separate segments (m/z 50-200 and m/z 200-1,000), allowing an independent tuning of the ion transmission parameters for both mass ranges. Three different implementations, involving either scanning from m/z 50-1000 in a single scan, or scanning from m/z 50-200 and from m/z 200-1000 in two alternating scans, or performing two separate HPLC-MS runs with m/z 50-200 and m/z 200-1000 scan ranges were critically assessed. The detected features were subjected to rigorous background filtering and quality control in order to obtain reliable metabolite features for subsequent differential quantification. The most efficient approach in terms of feature number, which forms the basis for statistical analysis, identification, and for generating biological hypotheses, was the separate analysis of two different mass ranges. This lead to an increase in the number of detectable metabolite features, especially in the higher mass range (m/z greater than 400), by 2.5 (negative mode) to 6-fold (positive mode) as compared to analysis involving a single scan range. The total number of features confidently detectable was 560 in positive ion mode, and 436 in negative ion mode.