Nearly 99 percent alike in genetic makeup, chimpanzees and humans might be even more similar were it not for what researchers call "lifestyle" changes in the 6 million years that separate us from a common ancestor. Specifically, two key differences are how humans and chimps perceive smells and what we eat.

A massive gene-comparison project involving two Cornell University scientists, and reported in the latest issue of the journal Science (Dec. 12, 2003), found these and many other differences in a search for evidence of accelerated evolution and positive selection in the genetic history of humans and chimps.

In the most comprehensive comparison to date of the genetic differences between two primates, the genomic analysts found evidence of positive selection in genes involved in olfaction, or the ability to sense and process information about odors. "Human and chimpanzee sequences are so similar, we were not sure that this kind of analysis would be informative," says evolutionary geneticist Andrew G. Clark, Cornell professor of molecular biology and genetics. "But we found hundreds of genes showing a pattern of sequence change consistent with adaptive evolution occurring in human ancestors." Those genes are involved in the sense of smell, in digestion, in long-bone growth, in hairiness and in hearing. "It is a treasure-trove of ideas to test by more careful comparison of human and chimpanzee development and physiology," Clark says.

The DNA sequencing of the chimpanzee was performed by Celera Genomics, in Rockville, Md., as part of a larger study of human variation headed by company researchers Michele Cargill and Mark Adams.

Celera generated some 18 million DNA sequence "reads," or about two-thirds as many as were required for the first sequencing of the human genome. Statistical modeling and computation was done by Clark and by Rasmus Nielsen, a Cornell assistant professor of biological statistics and computational biology. Some of the analysis, which also compared the mouse genome, used the supercomputer cluster at the Cornell Theory Center. Clark explains, "By lining up the human and chimpanzee gene sequences with those of the mouse, we thought we might be able to find genes that are evolving especially quickly in humans. In a sense, this method asks: What are the genes that make us human? Or rather, what genes were selected by natural selection to result in differences between humans and chimps?" The study started with almost 23,000 genes, but this number fell to 7,645 because of the need to be sure that the right human, chimp and mouse genes were aligned.

According to Clark, all mammals have an extensive repertoire of olfactory receptors, genes that allow specific recognition of the smell of different substances. "The signature of positive selection is very strong in both humans and chimps for tuning the sense of smell, probably because of its importance in finding food and perhaps mates," says Clark. In addition to the great departure in smell perception, differences in amino acid metabolism also seem to affect chimps' and humans' abilities to digest dietary protein and could date back to the time when early humans began eating more meat, Clark speculates. Anthropologists believe that this occurred around 2 million years ago, in concert with a major climate change.

"This study also gives tantalizing clues to an even more complex difference -- the ability to speak and understand language," Clark says. "Perhaps some of the genes that enable humans to understand speech work not only in the brain, but also are involved in hearing." Evidence for this came from a particularly strong sign of selection acting on the gene that codes for an obscure protein in the tectorial membrane of the inner ear. One form of congenital deafness in humans is caused by mutations to this gene, called alpha tectorin.

Mutations in alpha tectorin result in poor frequency response of the ear, making it hard to understand speech. "It's something like replacing the soundboard of a Stradivarius violin with a piece of plywood," Clark notes. The large divergence between humans and chimps in alpha tectorin, he says, could imply that humans needed to tune the protein for specific attributes of their sense of hearing. This leads Clark to wonder whether one of the difficulties in training chimpanzees to understand human speech is that their hearing is not quite up to the task. Although studies of chimpanzee hearing have been done, detailed tests of their transient response have not been carried out.

Clark emphasizes that a study like this cannot prove that the biology of humans and chimps differ because of this or that particular gene. "But it generates many hypotheses that can be tested to yield insight into exactly why only 1 percent in DNA sequence difference makes us such different beasts," he says.

Also collaborating in the study were researchers at Applied Biosystems (Foster City, Calif.), Celera Diagnostics (Alameda, Calif.) and Case Western Reserve University in Cleveland. The Science report is titled, "Inferring non-neutral evolution from human-chimp-mouse orthologous gene trios."