Automating the Design of Digital Microfluidic Devices for Synthetic Biology

Digital microfluidic devices manipulate and process fluids as individual droplets on an array of electrodes. The benefits of using such devices stem from microfluidics, as well as engineering. On the microfluidics side, the miniaturization of fluids leads to reduced reagent consumption and faster time-to-result. On the engineering side, the device is (a) compact and portable as the electrode array can be actuated directly off battery with no need for additional pressure or vacuum pumps; (b) reconfigurable as the electrodes can be used interchangeably and (c) programmable as the droplet movement can be programmed directly from the computer.

However, digital microfluidic devices have not been used yet in large scale in synthetic biology. Indeed, the use of a universal system (i.e., a single generic device that can be re-programed for each new bio-protocol) rises contamination issues, while a disposable-cartridge system comes with the difficulty of customized design.

We developed several algorithms to automate the design process of digital microfluidic cartridges. The main advantage of using automation over human design is trade-off analysis: the algorithms are better at balancing two or more objectives, while humans can optimize mainly for one, e.g., currently, researchers design new devices that reduce either the costs or the time-to-result, but have a hard time designing for both objectives.

Our tool bridges the design expertise requirements and supports the interaction between human editing and automatic design. Our algorithm optimizes for fabrication cost, reagent cost and time-to-result at the same time. Concretely, starting from a bio-protocol, our algorithm obtains a continuous solution space that allows selection among trade-off alternatives.

We tested our tool on seven bio-protocols commonly used in synthetic biology. The obtained designs were fabricated as real digital microfluidic cartridges and successfully tested for functionality.

By providing an automated way to efficiently generate digital microfluidic devices, we advance their use in synthetic biology research. New specific devices can be automatically generated for any new procedure that is established. Moreover, we believe that the use of our automated tool can expand beyond laboratory research, towards personal use.

Thus, our tool provides a step towards a future in which even non-technical users will be able to create microfluidic devices for their personal applications, thereby democratizing parts of health care. Our future work is focused on exploring the use of a digital microfluidic devices as personal laboratories operated at home.