Abstract
Despite instrument and algorithmic improvements, the untargeted and accurate assignment of metabolites remains an unsolved problem in metabolomics. New assignment methods such as our SMIRFE algorithm can assign elemental molecular formulas to observed spectral features in a highly untargeted manner without orthogonal information from tandem MS or chromatography. However, for many lipidomics applications, it is necessary to know at least the lipid category or class that is associated with a detected spectral feature to derive a biochemical interpretation. Our goal is to develop a method for robustly classifying elemental molecular formula assignments into lipid categories for an application to SMIRFE-generated assignments. Using a Random Forest machine learning approach, we developed a method that can predict lipid category and class from SMIRFE non-adducted molecular formula assignments. Our methods achieve high average predictive accuracy (>90%) and precision (>83%) across all eight of the lipid categories in the LIPIDMAPS database. Classification performance was evaluated using sets of theoretical, data-derived, and artifactual molecular formulas. Our methods enable the lipid classification of non-adducted molecular formula assignments generated by SMIRFE without orthogonal information, facilitating the biochemical interpretation of untargeted lipidomics experiments. This lipid classification appears insufficient for validating single-spectrum assignments, but could be useful in cross-spectrum assignment validation.
Highlights
Lipidomics is the subdiscipline of metabolomics concerned with the analytical investigation of the lipidome, the set of lipid metabolites within a sample, and their roles within the metabolome.Unlike other categories of metabolites that are largely grouped based on their structures, lipids are defined by their very low solubility in water and collectively represent a structurally and chemically diverse set of metabolites with various roles in normal and pathological cellular function
Our methods enable the lipid classification of non-adducted molecular formula assignments generated by SMIRFE without orthogonal information, facilitating the biochemical interpretation of untargeted lipidomics experiments
We present a novel method of predicting lipid category and class from molecular formula and information readily available from direct-infusion MS1 spectra without reliance on existing metabolite databases or orthogonal information
Summary
Unlike other categories of metabolites that are largely grouped based on their structures, lipids are defined by their very low solubility in water and collectively represent a structurally and chemically diverse set of metabolites with various roles in normal and pathological cellular function. The potential benefits of lipidomics are enormous, the rigorous analytical investigation of the lipidome in real-world biological samples requires the reliable observation of lipid features in the samples as well as the accurate assignment of those features to a lipid structure and/or lipid class This represents a significant bioanalytical chemistry problem due to the high structural diversity of lipids, their wide range of observed concentrations, and differences in lipid profiles across compartments and time [14,15,16]. Due to its sensitivity to a wide range of chemical structures and low detection limits, mass spectrometry remains the most popular analytical technique for lipidomics analysis [17]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
