Resolving Spectral Complexity in 4D Lipidomics Using a Two-Dimensional Deconvolution Framework.

The application of data-independent acquisition (DIA) in 4D lipidomics has been constrained by spectral interference due to fragment ion overlap, a bottleneck that existing one-dimensional deconvolution methods fail to fully resolve. Here, we overcome this limitation by introducing a two-dimensional liquid chromatography-ion mobility (LC-IM) deconvolution framework that unlocks the full potential of 4D lipidomics. By mathematically modeling the orthogonal LC-IM separation dimensions, our method reconstructs high-quality MS/MS spectra from highly complex DIA data, effectively disentangling co-eluting lipid interferences. We demonstrate the power of this approach by annotating 491 lipids from 1 µL human plasma at a 1% false discovery rate, a two-fold increase in coverage compared to traditional methods. Beyond bulk analysis, we showcase its unique capability for spatial lipidomics, enabling deep profiling of laser-microdissected tissue regions equivalent to only hundreds of cells, revealing metabolic reprogramming in human hepatocellular carcinoma. We further integrate this workflow with six-plex isobaric labeling to achieve high-throughput, high-accuracy quantification in spatial tissue mapping. This transition from one- to two-dimensional deconvolution establishes a robust, sensitive platform for deep lipidome characterization, bridging the gap between proteomics-grade throughput and lipidomic structural complexity.