MIT researchers have automated experiments that tune the size of lipid nanoparticles (LNPs), fat-based carriers used to package RNA. An in-line instrument measures particles during production, and software uses the readings to adjust conditions in later runs.
The system first mixes lipids dissolved in ethanol with RNA in an acidic buffer. After a set delay, a second mixer adds more buffer and stops particle growth. Changing the delay, buffer concentration, or lipid mixture changes the particles’ final diameter.
A downstream light-scattering instrument estimates average particle size and its distribution every 7.5 seconds. The software pairs those readings with pump settings and can run planned experiments, parameter sweeps, or Bayesian optimization—a search method that uses earlier measurements to choose conditions for a target size. Bubbles, unstable flow, and contamination can make the readings noisy; the supplement says some runs remained noisy despite cleaning and data smoothing.
One planned experiment tested 45 combinations of three mixing delays, three buffer concentrations, and five lipid ratios. The optimization experiments used yeast RNA as a lower-cost stand-in for mRNA. The work addresses manufacturing and particle measurements, not whether a therapeutic reaches a target tissue or improves treatment outcomes.
Sources
- ACS Nano paper: “Autonomous Lipid Nanoparticle Engineering”
- Supplementary methods and data
- Image: Michael D. Buschmann, Manuel J. Carrasco, Suman Alishetty, Mikell Paige, Mohamad G. Alameh, and Drew Weissman, “Nanomaterial Delivery Systems for mRNA Vaccines” (2021), via Wikimedia Commons, unmodified, CC BY 4.0. This is a general LNP illustration, not the MIT platform.
