Abstract
For large-scale and long-term metabolomics studies that involve a large batch or multiple batches of analyses, batch effects cause nonbiological systematic biases that may lead to false positive or false negative findings. Quantitative monitoring and correction of batch effects is critical to the development of reproducible and robust metabolomics platforms either for untargeted or targeted analyses. To achieve sufficient retention and separation of a broad range of metabolites with diverse chemical structures and physicochemical properties, LC-MS/MS based targeted metabolomics often involves 3 complemented chromatographic separation methods, including reversed-phase liquid chromatography (RP-LC), hydrophilic interaction liquid chromatography (HILIC), and ion-pair liquid chromatography (IP-LC). The purpose of this study is to quantitatively evaluate intra-batch variations or injection order effects of the RP-LC, HILIC, and IP-LC methods for targeted metabolomics analyses, and develop strategies to minimize intra-batch variations and correct injection order effects for problematic metabolites. Both RP-LC and HILIC methods exhibit robust intra-batch reproducibility in 0.2 µM standard mix QC, with ∼96 % of the measured metabolites showing acceptable intra-batch variations (<20 %); whereas, the intra-batch reproducibility for some metabolites in cell matrix QC may be compromised due to stability issue, suboptimal chromatographic retention, and/or matrix effects causing ionization suppression and/or retention instability. The IP-LC method exhibits significant injection order effects, which could be effectively corrected by the developed exponential models of signal drift trends as a function of injection order for individual targeted metabolites.