How many molecular changes does it take to turn a chimpanzee into a human being? Justin Fay, a geneticist with Lawrence Berkeley National Laboratory (Berkeley Lab), can't tell you just yet—but he can tell you that Charles Darwin's evolutionary engine, the process of natural selection, does reach down to the molecular level.

"We find strong evidence against the neutral theory of molecular evolution when we look at DNA variations across entire genomes rather than within single genes," Fay says. "Our findings suggest that rather than being driven by mutation and drift, the molecular evolution has been shaped by positive Darwinian selection."

In a paper published in the journal Nature on February 28, 2002, Fay, who is now with the Genome Sciences Department in Berkeley Lab's Life Sciences Division, along with Gerald Wyckoff and Chung-I Wu of the University of Chicago, reported on a study in which they examined data from 45 different gene surveys and compared "polymorphism" in the common fruit fly, Drosophila melanogaster, with "divergence" from its close cousin, Drosophila simulans. Polymorphism is the difference in DNA sequences among individual members of the same species, and divergence is the variation in DNA sequences between different species.

Justin Fay studied DNA variations across entire genomes of fruit flies and found evidence that natural selection drives mutations at the molecular level.

"Our comparisons of Drosophila melanogaster with Drosophila simulans found that the A/S divergence ratio was twice as high as the A/S polymorphism ratio," Fay says. "A higher A/S ratio of divergence has also been observed in other species comparisons, which suggests a rate of adaptive evolution in molecules that is far greater than permitted by the neutral theory."

That changes in the phenotypes (physical characteristics) of organisms selectively occur to make them more adaptive to their environment can readily be observed, as Darwin showed. However, since 1968, when the Japanese geneticist Motoo Kimura proposed the neutral theory of molecular evolution, there has been a scientific debate over whether natural selection plays much of a role in genotype changes. Kimura's theory held that at the level of DNA or amino acid sequences, most changes are neutral.

If the neutral theory is correct, then changes in DNA and amino acid sequences should be relatively constant, and the amount of polymorphism within a single species should be proportional to the amount of divergence between two different species. This means there should be a "molecular clock" of protein evolution that could be used to date the divergence between different species by comparing differences in their DNA and amino acid sequences. The neutral theory agreed well with experimental results that involved single gene studies, but the molecular clock was not always reliable.

Fay and his colleagues reconciled the contradictions by comparing the composite pattern from a large number of genes across the genomes of different fruit fly species, rather than single genes. They also took the important step of discarding "deleterious" polymorphisms, those mutations that negatively affect an organism and would eventually be eliminated from populations by natural selection.

"The effects of positive selection can be obscured by deleterious mutations that inflate the A/S ratio of polymorphism but not divergence," explains Fay. "Removal of the deleterious changes allowed us to make a clear comparison between neutral and adaptive changes."

For their polymorphism data, Fay and his colleagues focused on "common frequency mutations," which they defined as those found in more than 12.5 percent of their population samples. Divergence data were obtained by comparing a randomly chosen DNA sequence of D. melanogaster with that of D. simulans or, if unavailable, that of D. mauritana or D. sechilla.

"In going from gene studies to genomic studies, we found a substantial amount of positive selection taking place in protein evolution," Fay says. "Most neutral theory supporters have suspected that this positive selection was happening but they required a high standard of evidence to prove it. We've provided them with that evidence."

The next step, according to Fay, is to examine the DNA sequences that do not code for proteins. In the human genome, for example, only about one percent of DNA sequences actually carry instructions for making proteins. The remainder includes sequences that regulate the expression of genes—turns them on or off. Little has been done to determine the role mutations in these sequences may play in evolution. Fay is now part of a collaboration in the laboratory of Michael Eisen of the Life Sciences Division, investigating the evolution of complex phenotypes in different species of yeasts.

"Maybe phenotype variations in organisms arise as a result of changes in gene expression," he speculates. "Right now we only have a few examples of molecular changes that have been responsible for changes in a phenotype but this area hasn't really been explored and there's a lot of potential for new findings."

Additional information:

• More about the Eisen Lab's research on gene expression