Obtaining quantitative LNP data requires a technique that can directly visualize individual nanoparticles. Cryo-electron microscopy (cryo-EM) provides exactly that capability, imaging LNPs in a vitrified, near-native state that preserves their true structure [6]. While the imaging itself is well established, analyzing the micrographs remains a major bottleneck in traditional workflows: researchers manually inspect the micrographs and measure particle properties (such as size and morphology) one particle at a time [7]. Because this process is slow, manual or even semi-automatic picking of 100,000 particle images would require several man-months [8]. Hence, the limitation of traditional workflows is therefore not the imaging itself, but the manual analysis that follows. AI-powered platforms remove this bottleneck by automating particle annotation across entire datasets. Vitrification and imaging remain unchanged; the transformation lies in annotation: ATEM’s cryo-EM platform segments every detectable LNP across a full dataset, classifying each into Solid Core, Biphasic Split or Biphasic Dense, while simultaneously extracting per-particle size, aspect ratio, lamellarity and blebbing. This way, thousands of nanoparticles are annotated in seconds rather than months.

In consequence, the impact is not just speed, but scale. By analyzing thousands of particles per dataset, researchers turn cryo-EM images into statistically meaningful and precisely reproducible measurements of LNP morphology. Analyzing the full dataset ensures that even low-frequency morphology classes are captured, rather than lost in a small manual subset. In practice, this means researchers can directly quantify how particle population changes, including how specific morphologies convert into others [9].