Fig. 3

Overview of overlaying and combining MALDI-MSI and spatial transcriptomics data on human glioblastoma tissue. (A) Shows the schematic workflow used using ESCDAT. First MALDI-MSI (timsTOF fleX) and spatial transcriptomics (Xenium) are sequentially conducted on the same tissue section. Afterwards, using ESCDAT, a low mass resolution dataset is spatially aligned with the spatial transcriptomics data. Using fiducial markers and visible laser spots in the tissue generated via the MALDI laser. Using the cell boundaries from the cell segmentation staining, pixels per cell are averaged and a high-resolution m/z spectrum is created per cell. These data modalities are afterwards combined, and downstream analysis are performed using Seurat V5 and Python. (B) Visualization of co-registration MALDI-MSI with spatial transcriptomics. (I) A representative m/z ion map from MALDI-MSI showing the spatial distribution of a selected molecule across the tissue. (II) Fluorescence image of the same section with labelled markers as reference. (III-IV) Zoom into the red-boxed region: (III) MALDI-MSI view of a sharp fiducial marker corner; (IV) matching fluorescence view with the laser-ablation dots used as fiducials. (V) Overlay of MALDI sampling points (red circles) on the fluorescence image after co-registration. (VI) Final overlay of MALDI-MSI data with cell-boundary segmentation from spatial transcriptomics, demonstrating precise multimodal alignment. (VII) Direct overlay of fluorescence fiducials and MALD-MSI intensity map. The MALDI signal recapitulates the per-pixel fluorescence pattern, confirming grid-anchored, pixel-scale alignment in fully gridded regions.