Computational analysis and ratiometric comparison approaches aimed to assist column selection in hydrophilic interaction liquid chromatography-tandem mass spectrometry targeted metabolomics.

In the present work two different approaches, a semi-quantitative and a Derringer function approach, were developed to assist column selection for method development in targeted metabolomics. These approaches were applied in the performance assessment of three HILIC columns with different chemistries (an amide, a diol and a zwitterionic phase). This was the first step for the development of a HILIC UPLC-MS/MS method that should be capable to analyze a large number of polar metabolites. Two gradient elution profiles and two mobile phase pH values were tested for the analysis of multi-analyte mixtures. Acquired chromatographic data were firstly treated by a ratiometric, "semi-quantitative" approach which quantifies various overall analysis parameters (e.g. the percent of detected compounds, retentivity and resolved critical pairs). These parameters were used to assess chromatographic performance in a rather conventional/traditional and cumbersome/labor-intensive way. Secondly, a comprehensive and automated comparison of the three columns was performed by monitoring several well-known chromatographic parameters (peak width, resolution, tailing factor, etc.) using a lab-built programming script which calculates overall desirability utilizing Derringer functions. Derringer functions exhibit the advantage that column performance is ultimately expressed in an objective single and quantitative value which can be easily interpreted. In summary, results show that each column exhibits unique strengths in metabolic profiling of polar compounds. The applied methodology proved useful for the selection of the most effective chromatographic system during method development for LC-MS/MS targeted metabolomics, while it could further assist in the selection of chromatographic conditions for the development of multi-analyte methods.