A new machine that sorts fly and worm embryos could open the door to investigations of some human diseases, its inventors predict. Matthew Scott and his colleagues at Stanford University in California have devised a technique to separate useful mutant embryos from their normal siblings at the rate of 15 per second1.

The genes carried by mutant embryos can cause them to mimic human conditions, including nerve and heart diseases, and so are of great interest to researchers.

Embryos of the Drosophila fruitfly and the parasitic worm Caenorhabditis elegans develop fully within 24 hours. Selecting useful subjects by hand (often with a paintbrush) wastes time, and restricts the scope and number of experiments that researchers can perform.

The new machine, which sorts embryos according to their fluorescence, is 99% accurate, Scott says. "Any lab with skilled technicians should be able to build one and we would be happy to help them," he says. His team has posted the instructions for constructing the device on the Internet see: www.nature.com/nbt/web_extras/.

"It sounds impressive," says Allan Spradling, a developmental biologist at Johns Hopkins University in Baltimore, Maryland. "I think this is something that researchers in Drosophila early embryos have been wishing they had for a long time." Manual sorting of embryos is tedious, he adds, and limits both the design and execution of experiments.

Many of the genes that control fly and worm functions (the heart, for example) are astonishingly similar to those in people. So knocking out these genes and studying how the fly and worm embryos then develop is very useful. But such experiments come with two strings attached.

First, the organisms carry two copies of each gene, and often both must be corrupt to produce an effect in the developing embryo. But, second, flies carrying two corrupt genes have their short lives truncated even further, often not even surviving long enough to breed.

Researchers ensure the legacy of their genetic engineering by maintaining populations of flies carrying one good gene and one bad one. Breed these flies, and one-quarter of the resulting embryos get bad copies from both mum and dad -- and the experience of being peered at through a microscope by interested biologists.

But three-quarters of the embryos are disappointingly healthy and so are disposed of. This is where Scott's sorter comes in. To help spot which embryos have what genes, researchers 'mark' healthy copies of the gene with another gene that makes the embryo fluoresce under laser light.

The new device uses a laser to scan embryos suspended in solution and, by means of gates controlled by electromagnets, sorts the dim from the bright.