For a long time, genes were just an abstract concept, a name for whatever it was that carried inherited traits from one generation to the next. Now we know that the physical embodiment of a gene is a sequence of bases in a cell's DNA that, in turn, codes for the sequence of amino acids in a protein. But to make effective use of that knowledge, researchers have to solve the next part of the puzzle: How does the protein finds its final, biologically active shape, and just what does that shape mean?

Anyone who reads cereal boxes knows that skin, hair and muscles are made of proteins. But so are enzymes, hemoglobin, hormones and neurotransmitters. The walls of cells are a potpourri of proteins that receive signals from hormones or neurotransmitters and control what gets into and out of the cell. To make a protein, a cell strings amino acids together following the code in a gene. Almost immediately the chain folds into a complex shape, somewhat the way a piece of string will knot up if you twist the ends in opposite directions.

For a protein, function follows form. The proteins in skin, hair and muscles must have shapes that link together in a working structure. Enzymes must fit precisely onto the molecules whose reactions they control. Hormones must fit the receptors on the outside of cells, and the receptors must be the right shapes to accept hormones and other chemicals that interact with the cell.

The genetic code tells us the sequence of amino acids in a protein. We can read the code, but we don't always understand what we have read. To do that we have to find the folded shape of the protein and what function that shape represents. Among the researchers at Cornell pursuing this problem are Harold Scheraga and David Shalloway, using laboratory experiment and computer simulation on the Cornell Theory Center's IBM SP2 parallel-processing supercomputer to try to understand and predict how folding proceeds, and Steven Ealick, using X-ray crystallography at MacCHESS to find the structure of completely folded proteins. Follow the links for brief accounts of their work, and see also The Pathway from DNA to Drugs.

Three different ways of looking at a protein X-ray crystallography yields a wire-frame diagram (left) of electron densities. Knowing the sequence of amino acids, researchers insert them by hand in a 3-D graphics program. The computer can then create many different images. A ribbon drawing (center) reveals often-repeated structures such as an alpha helix or beta sheet. A "space-filling" image (right) shows each atom as a sphere, useful for studying the ways proteins bind to one another.