Spatially resolved metabolomics is an excellent tool for elucidating in situ molecular events, but its use remains challenging due to the complexity of the endogenous metabolites in bio-tissue and tissue heterogeneity. In this study, a data processing pipeline for spatially resolved metabolomics analysis of tumor microregion heterogeneity was developed and built into a graphical interface with MSI software. Biological tissue sections were analysed by ambient air-flow assisted desorption electrospray ionization mass spectrometry imaging. Histology-driven and characterized ion images overlay combined with metabolic feature-based spatial segmentation were developed to accurately extract the metabolic profile from the tissue microregion of interest. In addition, appropriate data pretreatment methods were investigated to evaluate their ability to identify biological variations from the complicated spatially resolved metabolomics data. Diverse graphical metabolic feature extraction and various data pretreatment methods enable not only the achievement of the best multivariate statistical results in an intuitive and simple way but also the discovery of low-abundance but reliable biomarkers. The results from a papillary thyroid cancer tissue study demonstrated that this data processing pipeline is a powerful and easy-to-use tool for investigating the spatial molecular events in tumor microenvironments and to therefore thoroughly understand their metabolic heterogeneity.
Keywords: Air-flow assisted ionization; Data processing; Mass spectrometry imaging; Spatially resolved metabolomics; Tumor heterogeneity.
Copyright © 2019 Elsevier B.V. All rights reserved.