This is a legitimately amazing paper. TL;DR: They made a microfluidic chip that can trap a million individual cells, hooked it up to a microscope, and used it to connect genetic circuits --> single-cell behaviors. Say you want to build a complicated gene circuit, like an oscillator with a period of six hours. Normally, you’d make a design, put it into cells, watch it fail, redesign it, and try again. With this system, you can instead make a million variants in a single tube, introduce them into cells, and trap those cells in the chip to see which designs work. You can then pull out a “winning” cell using optical tweezers, or identify which gene circuit it received via in situ sequencing. (Each circuit carries a short genetic barcode. You sequence the gene circuit variants beforehand to link each barcode to a design. Then, reading those barcodes inside the chip, using in situ sequencing, tells you which circuit each cell received.) This is a much more scalable way to test and optimize complex ideas in synthetic biology; not only gene circuits, but perhaps also metabolic pathways. I suspect it’ll be super useful for collecting data for predictive AI models, too. Congrats @NoahOlsman!