Ultrahigh-throughput Chemical and Biological Screening with Microfluidic Droplets

Many important targets for directed evolution, like enzymes or biosynthetic pathways, yield molecules that are untethered to the proteins and genes responsible for producing them. Conventionally, evolving targets like these has been accomplished in microtiter plates or, when possible, in live cells via selections. While microtiter plates are amenable to assays that cannot be performed in cells, they are expensive and limited in the size of the libraries they can screen. Conversely, selections can screen large libraries, but are only possible with targets that can be assayed in living cells. We are developing an ultrahigh-throughput screening technology that combines the throughput of selections with the flexibility and generality of microtiter plate screening. In our approach, picoliter aqueous droplets used as "reactors" to assay individual members of the target libraries. Due to the small size of the droplets, and the ability to form, process, and sort them at kilohertz rates, we are able to screen libraries numbering in the hundreds of millions using microliters of total reagent, reducing cost and increasing throughput compared to microtiter plates by 1 million-fold. In addition, by integrating microfluidic screening with robotic fluid handling, we will be able to perform completely automated, continuous directed evolution of these targets.