Using variable data independent acquisition for capillary electrophoresis-based untargeted metabolomics
Saki Kiuchi,Yasuhiro Otoguro,Tomoaki Nitta,Mi Hwa Chung,Taiki Nakaya,Yuki Matsuzawa,Katsuya Oobuchi,Kazunori Sasaki,Hiroyuki Yamamoto,Hiroshi Tsugawa
DOI: https://doi.org/10.1101/2024.01.17.576157
2024-01-20
Abstract:Capillary electrophoresis coupled with tandem mass spectrometry (CE-MS/MS) offers advantages in peak capacity and sensitivity for metabolic profiling, owing to the electroosmotic flow-based separation. However, the utilization of data-independent MS/MS acquisition (DIA) is restricted due to the absence of an optimal procedure for analytical chemistry and its related informatics framework. We assessed the mass spectral quality using two DIA techniques, namely, all-ion fragmentation (AIF) and variable DIA (vDIA), to isolate 60∼800 Da precursor ions with respect to annotation rates. Our findings indicate that vDIA, coupled with the updated MS-DIAL chromatogram deconvolution algorithm, yields higher spectral matching scores and annotation rates compared to AIF. Additionally, we evaluated a linear migration time (MT) correction method using internal standards to accurately align chromatographic peaks in a dataset. After the correction, the peaks exhibited less than 0.1 min MT drifts, a difference mostly equivalent to that of conventional reverse-phase liquid chromatography techniques. Moreover, we conducted MT prediction for metabolites recorded in mass spectral libraries and metabolite structure databases containing a total of 469,870 compounds, achieving an accuracy of less than 1.5 min root mean squares. Thus, our platform provides a peak annotation platform utilizing MT information, accurate precursor , and the MS/MS spectrum recommended by the metabolomics standards initiative. Applying this procedure, we investigated metabolic alterations in lipopolysaccharide (LPS)-induced macrophages, characterizing 170 metabolites. Furthermore, we assigned metabolite information to unannotated peaks using an structure elucidation tool, MS-FINDER. The results were integrated into the nodes in the molecular spectrum network based on the MS/MS similarity score. Consequently, we identified a significantly increased amount of metabolites in the LPS-administration group, glycinamide ribonucleotide, not present in any spectral libraries. Additionally, we retrieved metabolites of false-negative hits in the initial spectral annotation procedure. Overall, our study underscores the potential of CE-MS/MS with DIA and computational mass spectrometry techniques for metabolic profiling.
Biochemistry