To understand disease and develop new drugs, researchers often must begin by sorting the jumble of cell types in a living organism-tumor cells from normal cells, for instance. In some cases, a refrigerator-sized “fluorescence-activated cell sorter,” or FACS, can do the job. These machines, however, are expensive ($250,000), tricky to operate and prone to contamination. Now a team at Caltech, led by applied physicist Stephen Quake, has built a “microFACS,” reducing the complicated system of pumps, tubes and nozzles to micrometers-wide channels in a stamp-sized rubber chip.

George Whitesides, a Harvard University chemist who developed some of the technology used by the Caltech team says, “Quake brings the perspective of a physicist to this problem in bioanalysis, with results that are, to me, spectacular.” Bay Area-based Mycometrix aims to make prototype microFACS systems available to potential customers by year’s end; the startup was formed in 1999 to commercialize technologies from Quake’s lab. Vice president of business development Todd Krueger estimates that the reader will sell for $40,000, the disposable chips for less than $20 apiece. By making cell sorting cheaper and simpler, he says, the system should open up a host of new applications for the technique outside the lab, including doctor’s-office diagnosis and food or water screening.

The microFACS uses laser light and electricity to sort a cell mixture. Three 2-millimeter-wide wells on the chip are connected by a T-shaped network of channels. A researcher first tags the cells with fluorescent markers, then loads them into the bottom well and drives them toward the channel junction with an electric current. The channels narrow to just a few micrometers near the junction, forcing the cells to proceed in single file. A laser aimed at the junction hits the different markers and causes each cell type to fluoresce in a particular color-say, red for tumor and blue for normal. A detector prompts a computer to adjust the current to drive each cell to the left or the right, into a waste well or a collection well.

One reason the microFACS is so cheap is that the chips are made by soft lithography, an experimental process developed by Whitesides. It’s potentially far less expensive than micromachining; instead of carving individual chips, researchers cast them in a reusable silicon mold.

Whitesides calls Quake’s cell sorter “an elegant piece of design,” and points out that it could be coupled to other ultrasmall analytical systems, such as single-cell DNA sequencers. Indeed, one of Quake’s future aims is to build the sorter and all the other scaled-down devices needed for a particular experiment linked together on one chip, a goal that many others in the field are hotly pursuing.