[blackbear_]

This article [1] has a section with several references on that, with the most used technique [2] being depicted in fig. 3 (this is from 2016 so there might be a more recent technique around now).

The basic principle is to mix a solution with lipids and one with the mRNA and pump it through a channel with herringbone-shaped incisions that generate turbulence in the fluid. Apparently, the turbulence makes lipids surround the mRNA and stick together to form the nano-particle.

[1] http://www.future-science.com/doi/10.4155/tde-2016-0006

[2] https://science.sciencemag.org/content/295/5555/647.full

[neuronic]

So to quickly summarize they use fluid dynamics and force and whatever sticks through this process will be useful.

I wonder what the rate of lipid-surrounded mRNA is after the process. Say you fire 100 mRNAs through this construct, will 90 of them be surrounded by lipids afterwards? Or 30? Or 100?