A perfect prescription would fix what ails you, and leave the rest of you well enough alone. That's the ideal. The reality: Side effects bedevil almost all available drugs, and keep many others from ever reaching the market. A new approach for monitoring drugs' consequences in yeast cells could help sort the silver bullets from bombs more efficiently. Acacia Biosciences-a startup in Richmond, Calif.-is systematically "knocking out," or disabling, each yeast gene and studying what effect the loss of that specific gene has on the cell. The readout serves as a "molecular fingerprint of what the perfect drug would do," says Acacia CEO Bruce Cohen, since such a therapeutic agent would block the function of one-and only one-gene.

Acacia is working with Eli Lilly to study test compounds in this "genome reporter matrix." Researchers introduce candidate chemicals into thousands of yeast colonies, and record the compounds' influence on each gene. The drug that gets closest to the "knock-out" fingerprint might be put on the inside track toward human testing.