LipoGrid: A high-throughput multi-omics perturbation screen dissects the genetic architecture of lipid metabolism

Research Overview Lipids are essential components of cellular membranes, energy storage, and signaling, yet the genetic regulation of cellular lipid composition remains poorly understood. Here we introduce LipoGrid, a spatial mass spectrometry platform that links CRISPR/Cas9 perturbations to single-cell lipidomic profiles. By arraying genetically perturbed cells on micropatterned grids, LipoGrid quantified 158 lipid species across 143 gene knockouts in an internally controlled screen. Most perturbations produced distinct lipidomic signatures, often affecting specific lipid classes and molecular species. The platform recapitulated known gene–lipid relationships, including enzyme substrate specificities, pathway regulators, and disease-associated phenotypes, while also revealing previously unrecognized regulatory interactions. These results establish LipoGrid as a scalable approach for systematically mapping the genetic architecture of cellular lipid metabolism. Protocol Description Patterned cells were fixed in ice-cold 4% paraformaldehyde for 15 min at room temperature, washed with DPBS, and rinsed with ice-cold 150 mM ammonium formate (10 × 20 s, followed by 5 × 10 s; fresh solution each wash). Slides were dried under nitrogen for 15 min and scanned (FS120, Braun). Matrix deposition was performed using an HTX M5 sprayer with 7 mg/mL N-(1-naphthyl)ethylenediamine dihydrochloride (NEDC) in 70% methanol. Eighteen layers were applied in CC pattern (3 mm track spacing) at 75°C nozzle temperature, 1200 mm/min velocity, 10 psi nitrogen pressure, 0.60 mL/min methanol:water (1:1, v/v) pump flow, and 25°C plate temperature. Spatial lipidomics by MALDI-MSI MALDI-MSI was performed on a timsTOF fleX MALDI-2 mass spectrometer (Bruker Daltonik). Slides 1–2 were acquired in positive ion mode (m/z 300–1800), while slides 3–4 were acquired sequentially in positive and negative ion modes (m/z 300–2000). All acquisitions used a 26 × 26 µm laser beam with 30 µm pixel size and MALDI-2 post-ionization. Acquisition parameters differed between runs only in laser settings (40 shots, 1 kHz vs. 200 shots, 10 kHz), Deflection 1 delta (70 V vs. −70 V), and pre-pulse storage (5 vs. 10 µs). All other instrument settings were identical, including pre-TOF transfer time (85 µs), RF voltages, collision cell settings, and minimum quadrupole mass (m/z 300). Data were acquired with timsControl v6.1.5.0, converted to .imzML using SCiLS Lab v2025b Pro, and slides were stored at −80°C for up to 5 days before 10x Xenium analysis.