Earth's first life-form, floating in the proverbial froth of the primordial seas that eventually gave rise to trees, bees and humans, is not just a popular Darwinian conceit but also an essential biological premise that many researchers rely on as part of the foundation of their work.

In the 19th century, Charles Darwin went beyond others, who had proposed that there might be a common ancestor for all mammals or animals, and suggested that there was likely a common ancestor for all life on the planet—plant, animal and bacterial.

A new statistical analysis takes this assumption to the bench and finds that it not only holds water but indeed is overwhelmingly sound.

Was it not already obvious, from the discovery and deciphering of DNA, that all life forms are descended from a single common organism—or at least a basal species? No, says Douglas Theobald, an assistant professor of biochemistry of Brandeis University and author of the new study, detailed in the May 13 issue of Nature. (Scientific American is part of Nature Publishing Group.) In fact, he says, "When I went into it, I really didn't know what the answer would be."

Despite the difficulties of formally testing evolution—especially back across the eons to the emergence of life itself—Theobald was able to run rigorous statistical analyses on the amino acid sequences in 23 universally conserved proteins across the three major divisions of life (eukaryotes, bacteria and archaea). By plugging these sequences into various relational and evolutionary models, he found that a universal common ancestor is at least 10^2,860 more likely to have produced the modern-day protein sequence variances than even the next most probable scenario (involving multiple separate ancestors).*

"Evolution does well where it can be tested," says David Penny, a professor of theoretical biology at the Institute of Molecular BioSciences at Massey University in New Zealand and co-author of an accompanying editorial. Yet, he notes that evolution can make "testable predictions about the past (especially quantitative ones)" tricky at best. "That Theobald could devise a formal test," he says, "was excellent…. It will probably lead to a jump in what is expected of the formal evaluation of hypotheses, and that would help everybody."

Common ancestor acrimony

The mid-20th-century discoveries about the universality of DNA "really nailed it for people" in terms of establishing in popular—and academic—culture that there was a single universal common ancestor for all known life on Earth, Theobald says. And since then, "it's been widely assumed as true," he notes.

But in the past couple decades, new doubt has emerged in some circles. Microbiologists have gained a better understanding of genetic behavior of simple life forms, which can be much more amorphous than the typical, vertical transfer of genes from one generation to the next. The ability of microbes such as bacteria and viruses to exchange genes laterally among individuals—and even among species—changes some of the basic structural understanding of the map of evolution. With horizontal gene transfers, genetic signatures can move swiftly between branches, quickly turning a traditional tree into a tangled web. This dynamic "throws doubt on this tree of life model," Theobald says. And "once you throw doubt on that, it kind of throws doubt on common ancestry as well."

With the discovery of archaea as the third major domain of life—in addition to bacteria and eukaryotes—many microbiologists became more dubious of a single common ancestor across the board.

A test for evolution

Other researchers had put certain sections of life to the test, including a similar 1982 statistical analysis by Penny testing the relation of several vertebrate species. Theobald describes the paper as "cool, but the problem there is that they aren't testing universal ancestry." With advances in genetic analysis and statistical power, however, Theobald saw a way to create a more comprehensive test for all life.

In the course of his research, Theobald had been bumping against a common but "almost intractable evolutionary problem" in molecular biology. Many macromolecules, such as proteins, have similar three-dimensional structures but vastly different genetic sequences. The question that plagued him was: Were these similar structures examples of convergent evolution or evidence of common ancestry?

"All the classic evidence for common ancestry is qualitative and is based on shared similarities," Theobald says. He wanted to figure out whether focusing on those similarities was leading scientists astray.

Abandoned assumptions
Most people and even scientists operate under the premise that genetic similarities imply a common relation or ancestor. But as with similarities in physical appearance or structure, these assumptions "can be criticized," Theobald notes. Natural selection has provided numerous examples of convergent physical evolution, such as the prehensile tales of possums and spider monkeys or the long sticky insect-eating tongues of anteaters and armadillos. And with horizontal gene transfer on top of that, similar arguments could be made for genetic sequences.

"I really took a step back and tried to assume as little as possible in doing this analysis," Theobald says. He ran various statistical evolutionary models, including ones that took horizontal gene transfer into consideration and others that did not. And the models that accounted for horizontal gene transfer ended up providing the most statistical support for a universal common ancestor.

Murky origins
Theobald says his most surprising results were "how strongly they support common ancestry." Rather than being disappointed about simply backing up a long-held assumption, he says that at least, "it's always nice to know that we're on the right track."

These findings do not mean that a universal common ancestor establishes the "tree of life" pattern for early evolutionary dynamics. Nor, however, do they infer a "web of life" structure. The tree versus web debate remains "very controversial right now in evolutionary biology," Theobald says, reluctant to pick a side himself.

One of the other big unknowns remaining is just when this universal common ancestor lived and what it might have looked like—a question that will take more than Theobald's statistical models to answer. Theobald also notes that the support for a universal common ancestor does not rule out the idea that life emerged independently more than once. If other, fully distinct lineages did emerge, however, they either went extinct or remain as yet undiscovered.

Research will likely push on into these dusky corners of early evolution, Penny notes, as "scientists are never satisfied." He expects that researchers will try to sort back even earlier, before DNA took over, and assess the early stages of evolution during the RNA days.

On a more foundational level, Penny says, the paper should not put an end to the assessment of ancestral assumptions. Instead it should be a reminder that "we have never thought of all possible hypotheses," he says. "So we should never stop considering some new approach we haven't thought of yet."

*Erratum (5/13/10): This sentence was changed after publication. It originally stated that a universal common ancestor is more than 10 times more likely.