Enabling subcellular DESI-MSI for broad adoption: Acquisition, analysis, and application
- 1University of Pennsylvania
Abstract
Subcellular mass spectrometry imaging (MSI) has required specialized instrumentation and approaches. Additionally, data analysis has required expensive proprietary software, or code-based open-source tools. These barriers render spatial metabolomics less accessible than peer spatial omics techniques (transcriptomics, proteomics). Here we demonstrate subcellular MSI on standard commercial hardware, analyzed using code-free, open-source software. An inexpensive, fully reversible modification of a stock DESI source takes the platform past its typical 5-10 m pixels to routine 2x2 m pixels, and 1x1 m in proof-of-principle experiments. At a tissue-glass boundary imaged at 2x2 m, tissue-specific ions showed no measurable spillover onto the adjacent glass within the {+/-}4.4 m uncertainty of the edge position. To make these data interpretable, we extend MSI.EAGLE, an open-source, vendor-agnostic application with two new methods: i) Spatial UMAP preserves tissue architecture alongside subcellular detail, where conventional clustering breaks apart at high resolution; ii) Histology Fit performs landmark-free co-registration, and recovers synthetic misalignments of up to 8 m to within one pixel of a reference registration. Together, these enable previously high-burden spatial metabolomics experiments. In mouse brain tissue, we demonstrate genuine subcellular signal by resolving mitochondrial cardiolipin into perinuclear puncta, quantifying cytoplasmic enrichment across 6,073 cells. Cell identities are assigned to MSI pixels using histology-predicted transcriptomics, linking cell provenance to metabolism in 2D space directly. Finally, we recover cell-type-specific metabolomes consistent with the transcriptomic identities. This workflow allows cellular and subcellular spatial metabolomics linked to cell provenance, with only stock MSI hardware, open-source GUI-based software, and common histology techniques, putting it within reach of many researchers and laboratories.
Related
- Evolution and metabolite signaling of fumarases (class I and II): origin, function, and subcellular targeting
- Subcellular Localization and Cell Cycle-Associated Expression of Human dUTPase Isoforms
- STRAND: a comprehensive subcellular-resolved spatial transcriptome RNA architecture and navigation database
- A single domain intrabody biases the follicle-stimulating hormone receptor subcellular trafficking
- SpaCut: Transcript-Aware Morphology Fusion for Robust Cell Segmentation in Subcellular Spatial Transcriptomics.
BibTeX
@misc{boehmler2026enabling,
title = {{Enabling subcellular DESI-MSI for broad adoption: Acquisition, analysis, and application}},
author = {D. J. Boehmler and M. Loth and T. O'Harrow and G. Hoang and O. B. Patel and F. Akhtar and P. Kain and A. Sengupta and G. Paschos and M. Li and T. P. Gade and G. A. FitzGerald and A. M. Weljie},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.25.754519},
url = {https://www.biorxiv.org/content/10.64898/2026.09.25.754519v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Enabling subcellular DESI-MSI for broad adoption: Acquisition, analysis, and application
AU - D. J. Boehmler
AU - M. Loth
AU - T. O'Harrow
AU - G. Hoang
AU - O. B. Patel
AU - F. Akhtar
AU - P. Kain
AU - A. Sengupta
AU - G. Paschos
AU - M. Li
AU - T. P. Gade
AU - G. A. FitzGerald
AU - A. M. Weljie
PY - 2026
DA - 2026/09/25
PB - bioRxiv
DO - 10.64898/2026.09.25.754519
UR - https://www.biorxiv.org/content/10.64898/2026.09.25.754519v1
AB - Subcellular mass spectrometry imaging (MSI) has required specialized instrumentation and approaches. Additionally, data analysis has required expensive proprietary software, or code-based open-source tools. These barriers render spatial metabolomics less accessible than peer spatial omics techniques (transcriptomics, proteomics). Here we demonstrate subcellular MSI on standard commercial hardware, analyzed using code-free, open-source software. An inexpensive, fully reversible modification of a stock DESI source takes the platform past its typical 5-10 m pixels to routine 2x2 m pixels, and 1x1 m in proof-of-principle experiments. At a tissue-glass boundary imaged at 2x2 m, tissue-specific ions showed no measurable spillover onto the adjacent glass within the {+/-}4.4 m uncertainty of the edge position. To make these data interpretable, we extend MSI.EAGLE, an open-source, vendor-agnostic application with two new methods: i) Spatial UMAP preserves tissue architecture alongside subcellular detail, where conventional clustering breaks apart at high resolution; ii) Histology Fit performs landmark-free co-registration, and recovers synthetic misalignments of up to 8 m to within one pixel of a reference registration. Together, these enable previously high-burden spatial metabolomics experiments. In mouse brain tissue, we demonstrate genuine subcellular signal by resolving mitochondrial cardiolipin into perinuclear puncta, quantifying cytoplasmic enrichment across 6,073 cells. Cell identities are assigned to MSI pixels using histology-predicted transcriptomics, linking cell provenance to metabolism in 2D space directly. Finally, we recover cell-type-specific metabolomes consistent with the transcriptomic identities. This workflow allows cellular and subcellular spatial metabolomics linked to cell provenance, with only stock MSI hardware, open-source GUI-based software, and common histology techniques, putting it within reach of many researchers and laboratories.
ER -