MSI-ATLAS: Mass spectrometry imaging and explainable machine learning uncover the brain's lipid landscapes

Recent computational advances in mass-spectrometry imaging (MSI) now enable unprecedented insight into organ-wide molecular composition and functional architecture. Here, we present a first example of a high-resolution, molecular-computational atlas of specific mouse brain lipids - acquired using a NEDC matrix and negative mode MSI - covering 123 anatomically defined regions with 191 polygonal annotations derived solely from the MSI data -- no auxiliary imaging required. To overcome annotation ambiguity and MSI complexity, we introduced the Computational Brain Lipid Atlas (CBLA), a graph-based visual-explainability framework that generates virtual landscape visualizations (VLV) of specific lipids distributions across the brain substructures. The CBLA integrates dimensionality reduction and ensembles of supervised models to (i) refine annotations, (ii) elucidate interregional relationships, (iii) interpret model behavior, and (iv) formulate biologically testable hypotheses. The CBLA revealed novel lipid distribution patterns, functional integrations, anatomical connections - the brains telephone cables, and region-specific disease signatures - disease networks in the basal ganglia. A new algorithm decomposes annotated regions into precise m/z features and resolves full-precision m/z values from MSI data, producing a comprehensive high-resolution brain map. It can be used for any MS measurements: metabolites, lipids, and peptides. This resource underpins downstream studies, exemplified here by characterizing the lipid molecular composition of A{beta} plaques, their spatial arrangement, and their connections with the surrounding tissue. HighlightsO_LIMass Spectrometry Imaging (MSI) data were used to generate truthful visualizations for annotating brain regions with high resolution without the need for other modalities. C_LIO_LIMSI data were used to generate a computational atlas representation of the annotated brain regions. C_LIO_LIPathological structures reveal their origin and their effects on specific brain networks. C_LIO_LIAnatomical regions and functional networks show specific lipid patterns. C_LIO_LIBrain stem nuclei and the white matter have a unique composition of lipids revealing their involvement in pathological networks. C_LIO_LIThe atlas Virtual Landscape Visualizations (VLV) enable visualizing region-specific differences between different mouse models. C_LI